IEEE 802.11aa: Improvements on Video Transmission over Wireless

Alexander T.E.I. of Thessaloniki
School of Technological Applications
Department of Informatics
IEEE 802.11aa: Improvements on Video Transmission
over Wireless LANs
B.Sc. Thesis
by
Georgios Savvidis
Thesis Supervisor
Dr. Periklis Chatzimisios
Assistant Professor
Alexander T.E.I. of Thessaloniki
Department of Informatics
P.O. BOX 141
GR - 574 00, Thessaloniki
GREECE
May, 2012
ii
iii
IEEE 802.11aa: Improvements on Video Transmission
over Wireless LANs
B.Sc. Thesis
by
Georgios Savvidis
Alexander T.E.I. Of Thessaloniki
iv
v
vi
Abstract
In the recent years, audio and video applications have started to dominate the Internet.
Meanwhile, IEEE 802.11 networks have become a basic feature of mobile devices, the
popularity of which is growing fast alongside their technological capacities, providing the end
users almost the same capabilities as a Personal Computer. This results in users having
performance expectations as defined by their experience with wired video media.
While the current IEEE 802.11 standard provides high enough bandwidth for the needs of
video transmission, in addition with specifying mechanisms for a certain level of Quality of
Service (QoS), there are still several issues that need to be addressed in order to improve the
quality of a multimedia stream. In particular:

The legacy IEEE 802.11 does not provide mechanisms to ensure that all multicast
members successfully receive the transmitted data. In other words the current standard
does not provide the means for a reliable multicast/broadcast transmission.

It is essential to specify a method that mitigates the effects of Overlapping BSS
(OBSS) environments in order to offer increased robustness, without the need for
centralized management.

In the case of multiple video transport streams that belong in the same EDCA Access
Category, a mechanism is needed to provide prioritization between them.

In situations with insufficient channel capacity, there needs to be a mechanism that
marks the less important information as “discarded”, providing a more graceful video
degradation.

It is also important the applied mechanism for multimedia stream transport to be
compatible with the IEEE 802.1AVB protocol suite.
For all the above reasons, various researchers attempted to propose solutions and design
techniques until, finally, the IEEE decided to establish a working group, called the Task
Group aa (TGaa). This group aims at standardizing MAC layer enhancements for more
reliable multicast video transmission, while at the same time ensuring backwards
compatibility.
vii
viii
Acknowledgments
This thesis is the result of the past one year of a very interesting and valuable research at the
Alexander T.E.I. of Thessaloniki. Through this experience I had the opportunity to put in a
well-constructed frame all the relative knowledge that I gained during my B.Sc. studies.
However, nothing of this would have come true without the valuable help of some people that
I need to thank.
First of all, I would like to specially thank my thesis supervisor, mentor and friend, Dr.
Periklis Chatzimisios, Assistant Professor in the Department of Informatics at the Alexander
T.E.I. of Thesssaloniki, for assigning to me this particular thesis and giving me the chance to
work with and learn a lot of him, for his support and patience and for his guidance in matters
that regard the next step of my career.
Part of this work is credited to Kostas Maraslis, PhD candidate in the Department of
Telecommunications Science and Technology at the University of Peloponnese, whom I
sincerely thank for his kind cooperation and assistance during the whole process of this thesis'
implementation.
A deep thank to my old colleague and good friend, Georgios Papadopoulos for generously
sharing his knowledge whenever I needed his help.
I would also like to thank my professors Nikolaos Psarras and Stefanos Harhalakis from the
Department of Informatics at the Alexander T.E.I. of Thessaloniki, who imparted to me their
knowledge in the best possible way and helped me discover my interest in the field of
networking.
Finally, this thesis would not have been accomplished without the support of my family and
my very close friends who are always next to me.
Georgios Savvidis
ix
x
Contents
Abstract.....................................................................................................................................vii
Acknowledgments......................................................................................................................ix
Chapter 1.....................................................................................................................................1
Theoretical Background..............................................................................................................1
1.1 Introduction to IEEE.........................................................................................................1
1.2 The IEEE Wireless Zone..................................................................................................1
1.2.1 IEEE 802.11 WLAN.................................................................................................2
1.2.2 IEEE 802.15 WPAN.................................................................................................2
1.2.3 IEEE 802.16 WMAN................................................................................................3
1.2.4 IEEE 802.18 RR-TAG and IEEE 802.19 WC-TAG................................................4
1.2.5 IEEE 802.21 MIH.....................................................................................................5
1.2.6 IEEE 802.22 WRAN.................................................................................................5
Chapter 2.....................................................................................................................................7
IEEE 802.11 Standards Family...................................................................................................7
2.1 The IEEE 802.11 base standard........................................................................................7
2.1.1 Infrared (IR)..............................................................................................................7
2.1.2 Frequency Hopping Spread Spectrum (FHSS).........................................................8
2.1.3 Direct Sequence Spread Spectrum (DSSS)..............................................................9
2.1.3.1 Complementary Code Keying (CCK).............................................................10
2.1.4 Orthogonal Frequency-Division Multiplexing (OFDM)........................................10
2.2 IEEE 802.11 Amendments.............................................................................................11
2.2.1 IEEE 802.11a..........................................................................................................11
2.2.2 IEEE 802.11b..........................................................................................................11
2.2.3 IEEE 802.11d..........................................................................................................12
2.2.4 IEEE 802.11e..........................................................................................................12
2.2.5 IEEE 802.11f..........................................................................................................13
2.2.6 IEEE 802.11g..........................................................................................................14
2.2.7 IEEE 802.11h..........................................................................................................14
2.2.8 IEEE 802.11i...........................................................................................................15
2.2.9 IEEE 802.11j...........................................................................................................16
2.2.10 IEEE 802.11k........................................................................................................16
2.2.11 IEEE 802.11n........................................................................................................16
xi
2.2.12 IEEE 802.11p........................................................................................................18
2.2.13 IEEE 802.11r........................................................................................................18
2.2.14 IEEE 802.11s........................................................................................................19
2.2.15 IEEE 802.11w.......................................................................................................19
2.2.16 IEEE 802.11y........................................................................................................19
2.3 Ongoing standardizations...............................................................................................21
2.3.1 IEEE802.11ac.........................................................................................................21
2.3.2 IEEE 802.11ad........................................................................................................22
2.3.3 IEEE 802.11ae........................................................................................................22
2.3.4 IEEE 802.11af.........................................................................................................22
2.3.5 IEEE 802.11ai ........................................................................................................23
2.3.6 IEEE 802.11ah........................................................................................................23
Chapter 3...................................................................................................................................24
IEEE 802.11 Medium Access Control (MAC) Layer...............................................................24
3.1 Distributed Coordination Function (DCF).....................................................................25
3.1.1 Carrier Sense Multiple Access (CSMA).................................................................25
3.1.2 Collision Avoidance (CA), Random Backoff & Contention Window...................26
3.1.3 Request to Send/Clear to Send (RTS/CTS) Frame Exchange................................28
3.2 Point Coordination Function (PCF)................................................................................29
3.3 Hybrid Coordination Function (HCF)............................................................................30
3.3.1 HCF Contention Based Channel Access (EDCA)..................................................30
3.3.2 HCF Controlled Channel Access (HCCA).............................................................33
Chapter 4...................................................................................................................................35
Stream Classification Service and Interworking with 802.1AVB............................................35
4.1 Stream Classification Service.........................................................................................35
4.1.1 Intra-access category prioritization.........................................................................36
4.1.2 Graceful degradation...............................................................................................37
4.1.3 Stream Classification Service (SCS)......................................................................37
4.1.4 Related Work..........................................................................................................38
4.2 Interworking with 802.1AVB.........................................................................................39
4.2.1 IEEE 802.1AVB.....................................................................................................39
4.2.1.1 IEEE 802.1AS ................................................................................................41
4.2.1.2 IEEE 802.1Q...................................................................................................42
xii
4.2.1.2.1 IEEE 802.1Qat.........................................................................................43
4.2.1.2.2 IEEE 802.1Qav........................................................................................43
4.2.1.3 IEEE 802.1BA.................................................................................................44
4.2.2 IEEE 802.11aa: Interworking with 802.1AVB.......................................................44
Chapter 5...................................................................................................................................46
Multicast Issues and IEEE 802.11aa.........................................................................................46
5.1 MAC Layer Solutions.....................................................................................................46
5.1.1 Negative Feedback-Based Solutions......................................................................47
5.1.2 Positive Feedback-Based Solutions........................................................................48
5.2 Network Layer Solutions................................................................................................50
5.3 Transport Layer Solutions..............................................................................................52
5.4 Application Layer Solutions...........................................................................................52
5.5 Cross-Layer Solutions....................................................................................................55
5.5.1 Decentralized Networks (Ad-Hoc Networks).........................................................55
5.5.2 Centralized Networks (Access Point-Based Networks).........................................57
5.6 The IEEE 802.11aa Task Group Approach....................................................................61
5.6.1 Group Addressed Transmission Service (GATS)...................................................61
5.6.1.1 DMS Procedures.............................................................................................62
5.6.1.2 GCR Procedures..............................................................................................62
5.6.1.2.1 GCR Unsolicited Retry...........................................................................63
5.6.1.2.2 GCR Block Ack.......................................................................................63
5.6.2 Other Solutions.......................................................................................................65
Chapter 6...................................................................................................................................68
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa.........................................68
6.1 Industrial, Scientific and Medical (ISM) Radio Bands..................................................68
6.1.1 2.4 GHz Wireless Band..........................................................................................68
6.1.2 5 GHz Wireless Band.............................................................................................69
6.2 The Overlapping BSS Problem......................................................................................71
6.3 Importance of the OBSS Problem..................................................................................75
6.4 The IEEE 802.11aa Task Group Approach....................................................................78
6.4.1 QLoad Report Element...........................................................................................78
6.4.2 Calculating field values..........................................................................................80
6.4.3 Channel Selection...................................................................................................81
xiii
6.4.4 Sharing in an OBSS situation.................................................................................84
6.4.4.1 Proportional Sharing scheme [70]...................................................................85
6.4.4.2 On-demand Sharing scheme [70]....................................................................86
6.4.5 HCCA TXOP Negotiation......................................................................................87
6.5 Related Work..................................................................................................................92
6.5.1 Single channel management...................................................................................92
6.5.2 Dynamic channel switching....................................................................................93
6.5.3 Other solutions........................................................................................................94
Chapter 7...................................................................................................................................96
Conclusions and Further Research............................................................................................96
7.1 Conclusions....................................................................................................................96
7.2 Further Research.............................................................................................................96
List of Figures...........................................................................................................................98
List of Tables...........................................................................................................................100
Bibliography............................................................................................................................101
xiv
Chapter 1
Theoretical Background
1.1 Introduction to IEEE
The Institute of Electrical and Electronics Engineers (IEEE) was found in 1963 by the merger
of the American Institute of Electrical Engineers (AIEE) and the Institute of Radio Engineers
(IRE) [1]. IEEE is a non-profit leading organization for technology standards, research and
other professional and educational activities in the fields of Electronics, Electrical,
Telecommunications, Computer and Biomedical Engineering. As of today, it has more that
400.000 members worldwide.
The IEEE 802 project was requested in December 1979 [1], and its first meeting was held in
February 1980. Its object is the development of standards both for wired and wireless Local
and Metropolitan Area Networks (LANs/MANs). In the following sub-chapters there will be
presented a brief summary of the active IEEE 802 Working Groups (WG) in the wireless
zone.
1.2 The IEEE Wireless Zone
In the latest years wireless networks have expanded rapidly and nowadays they play a
significant role in business and home communication networks. More and more devices of
daily use, such as PDAs, laptops, mobile phones, etc, include wireless network support as a
basic feature. The term “wireless network”, however, is a very general term, since there are
various different types of networks that operate on the wireless medium. IEEE in order to
provide high quality support for all these kind of wireless networks, has deployed WGs,
where each of them has the responsibility of developing standards for a particular field.
Currently, the IEEE 802 project has the following active WGs, for the wireless zone:

IEEE 802.11 Wireless Local Area Network (WLAN)

IEEE 802.15 Wireless Personal Area Network (WPAN)

IEEE 802.16 Wireless Metropolitan Area Network (WMAN)

IEEE 802.18 Radio Regulatory – Technical Advisory Group (RR-TAG)

IEEE 802.19 Wireless Coexistence – Technical Advisory Group (WC-TAG)
1
Theoretical Background

IEEE 802.21 Media Independent Handoff (MIH)

IEEE 802.22 Wireless Regional Area Network (WRAN)
1.2.1
IEEE 802.11 WLAN
The first IEEE 802.11 standard was published in 1997. Since then numerous modifications
and enhancements have been made to the original standard, and today it is considered as one
of the most significant standardization achievements. The standard has a lot of similarities to
the IEEE 802.3 standard, that defines the PHY and MAC layers of wired Ethernet. IEEE
802.11 and its standards family are presented separately in Chapter 2.
1.2.2
IEEE 802.15 WPAN
The IEEE 802.15 WPAN focuses on the development of standards for Personal Area
Networks (PANs) or, in other words, for short distance wireless networks. The 802.15 WG
includes seven task groups:

IEEE 802.15.1 Bluetooth specification

IEEE 802.15.2 Coexistence between IEEE 802.15 and IEEE 802.11

IEEE 802.15.3 High-rate WPAN

IEEE 802.15.4 Low-rate WPAN

IEEE 802.15.5 Mesh Networking

IEEE 802.15.6 Body Area Network (BAN)

IEEE 802.15.7 Visible Light Communication (VLC)
IEEE 802.15.1 was published in June 2002. Defining PHY and MAC specifications, 802.15.1
comprises an adaptation of the Bluetooth Specification v1.1. It is also considered an additional
resource for Bluetooth devices constructors. For updating the 802.15.1 standard to the
Bluetooth v.1.2 specification, the task group 1a (TG1a) was formed [1]. Meanwhile, the IEEE
802.15.2 TG undertook the development of a recommended practice, aiming at facilitating the
coexistence of 802.11 and 802.15 networks.
IEEE 802.15.3 is a MAC and PHY standard that provides High-rate (20Mbit/s or greater),
among with low power and low cost solutions for the needs of portable consumer digital
imaging and multimedia applications, for WPANs [2]. 802.15.3 also provides a certain level
of QoS. On the other hand, IEEE 802.15.4 was developed to provide low data rate solutions
for WPANs. The emphasis is on very low complexity, among with extremely low power
2
Theoretical Background
consumption, that would allow battery duration from several months to several years. It is
mostly applied to applications such as sensors, remote controls, etc. IEEE 802.15.4 comprises
the basis for various specifications, with ZigBee being the most well-known between them.
Since neither 802.15.3, nor 802.15.4 define any path selection methods, TG5 was formed to
provide the means for mesh networking both for high and low data rate WPANs. The result of
this work was the IEEE 802.15.5 standard, which supports both full mesh and partial mesh
topologies. There are several advantages of mesh networking, such as the ability to extend the
network coverage without increasing the transmit power, the enhanced reliability via route
redundancy, easier network configuration and longer device battery life due to fewer
retransmissions [3].
In November 2007 the IEEE 802.15 TG6 was formed and began developing a low-power and
low-frequency communication standard optimized for devices and operation on, in or around
the human body, although it is not limited to humans. It serves a variety of applications
including medical, consumer electronics, and entertainment [4]. Finally, 802.15 includes the
TG7 the first meeting of which, was held in January 2009. IEEE 802.15.7 is an underdevelopment standard for visible light communications. It based on the idea of “what you see
is what you send”, it is aesthetically pleasing and since it operates on the visible light
spectrum, it is harmless for the human health. The standard is capable of delivering data rates
high enough to support both for indoor and outdoor audio/video multimedia applications.
1.2.3
IEEE 802.16 WMAN
The IEEE 802.16 working group was established in 1999 to develop standards and
recommended practices for broadband Wireless MANs. Though its official name is
WirelessMAN, it is mostly known as WiMAX (Worldwide Interoperability for Microwave
Access). IEEE 802.16 defines four different PHYs [1] supporting channel bandwidths of 1.25
MHz – 20 MHz, in any band in between the range 2 GHz – 66 GHz. It also includes an
adaptive modulation scheme that takes advantage of good signal conditions, Multiple-in
Multiple-out (MIMO) antennas and Hybrid Automatic Repeat Request (HARQ) for greater
error correction performance. The modulation scheme that it uses is Orthogonal FrequencyDivision Multiple Access (OFDMA), which is basically a multi-user version of the
Orthogonal Frequency-Division Multiplexing (OFDM).
3
Theoretical Background
The 802.16 MAC provides a means to encapsulate technologies of the wired medium like
Ethernet, Asynchronous Transfer Mode (ATM) and IP on the air interface. It also provides
security enhancements, by using both authentication and encryption mechanisms, such as the
Advanced Encryption Standard (AES) and the Data Encryption Standard (DES). Finally,
802.16 provides a strong QoS support, with 5 QoS classes.
1.2.4
IEEE 802.18 RR-TAG and IEEE 802.19 WC-TAG
IEEE 802.18 and 802.19 are considered Technical Advisory Groups. 802.18 does not develop
new standards but, instead it supports the IEEE 802 wireless working groups in radio
regulatory matters. Its main purpose is to monitor and to actively participate in ongoing radio
regulatory activities worldwide, while it is also the liaison to other standards of mutual interest
[5]. It currently supports the following six groups:

IEEE 802.11 Wireless Local Area Network (WLAN)

IEEE 802.15 Wireless Personal Area Network (WPAN)

IEEE 802.16 Wireless Metropolitan Area Network (WMAN)

IEEE 802.21 Media Independent Handoff (MIH)

IEEE 802.22 Wireless Regional Area Network (WRAN)
It also supports IEEE 802.20 Mobile Broadband Wireless Access (MBWA) which, however,
is in hibernation since March 2011 due to lack of activity.
IEEE 802.19 working group provides technical advise about coexistence to the following
working groups [5]:

IEEE 802.11 Wireless Local Area Network (WLAN)

IEEE 802.15 Wireless Personal Area Network (WPAN)

IEEE 802.16 Wireless Metropolitan Area Network (WMAN)

IEEE 802.22 Wireless Regional Area Network (WRAN)
With 802.19 it is assured that wireless devices operating in one of the above networks, will
not cause harmful interference to neighboring networks. To ensure that, sometimes 802.19 is
involved in the development of the mentioned standards. Furthermore, since January 2010, the
802.19 working group is developing a standard, the IEEE 802.19.1, that allows wireless
networks coexistence in the TV White Spaces.
4
Theoretical Background
1.2.5
IEEE 802.21 MIH
IEEE 802.21 standard, which was published in 2008, provides algorithms that enable
handover and interoperability both between networks of the same and of different types, that
may be either 802 or non-802 networks. Handover is defined by the European
Telecommunications Standards Institute (ETSI) and the 3rd Generation Partnership Project
(3GPP), as “the transfer of a user's connection from one radio channel to another”. Handovers
that happen within a single network are known as Horizontal Handovers, while those that are
triggers across different networks, are known as Vertical Handovers. 802.21 is mainly used
for Vertical Handovers, though it can be used for Horizontal Handovers, as well. Reasons for
triggering a handover include [1]:

Signal quality degradation.

Stronger signal reception from a neighboring cell.

The current signal strength has dropped under the minimum acceptable level.

The current signal quality has dropped under the minimum acceptable level.

The data traffic in the current cell has become high.
IEEE 802.21 achieves the seamless handover and interoperability between different types of
networks, by introducing a new, transparent layer between the layers 2 and 3 in the traditional
OSI/ISO model. This transparent layer is responsible for providing the information needed to
the upper layers, in order for the handover to be media independent [6]. Practical examples of
the use of IEEE 802.21 could be:

A user wishing to switch from an 802.3 network to an 802.11 one.

When a mobile phone user connected to a Global System for Mobile Communications
(GSM) network enters in range of an 802.11 network, he/she should be able to hand
off of the GSM and seamlessly connect to the 802.11.
1.2.6
IEEE 802.22 WRAN
The scope of IEEE 802.22 standard is to enhance the wireless interface, by modifying the
MAC and PHY layers, of point-to-multipoint WRANs of 40 km or more, operating in the
VHF/UHF TV broadcast bands between 54 MHz and 862 MHz [7]. The purpose of this
project was to fulfill the need of broadband wireless network access in areas where the wired
infrastructure is too expensive to be deployed. In such areas, the deployment of 802.22
operates in the spectrum that is allocated to the TV broadcast service, but is currently not
5
Theoretical Background
used. The major prerequisite of the standard was to be able to operate in this spectrum without
causing any harmful interference to the incumbent operation both in the VTF and the low
UHF TV bands.
The 802.22 working group was formed after the 802.18 study group decided that none of the
existing IEEE 802 PHY/MAC combination could achieve this goal without major
modifications. The 802.22 project was initially expected to be published in 2010, but it was
finally over in July 2011. The working group has two additional task groups, the TG1 and the
TG2, which are writing 802.22.1 and 802.22.2, respectively.
The 802.22.1 standard was developed to enhance harmful interference protection for low
power licensed devices operating in TV Broadcast Bands. The need of this standard arose
when the FCC proposed to allow new license-exempt devices to operate in the same channels
as the 802.22. Thus, it was important to protect devices, such as wireless microphones, that
were currently using this spectrum, from harmful interference. Finally, 802.22.2 is
Recommended Practice, currently under development, that describes the best engineering
practices and detailed technical guidance for the installation and deployment of IEEE 802.22
systems [7].
6
Chapter 2
IEEE 802.11 Standards Family
2.1 The IEEE 802.11 base standard
The initial version of the IEEE 802.11 standard was released in 1997 and specifies two data
rates of 1 and 2 Mbit/s. It specifies three alternative physical layer technologies:

Diffuse Infrared (IR)

Frequency Hopping Spread Spectrum (FHSS)

Direct Sequence Spread Spectrum (DSSS)
IEEE 802.11, also defined a Medium Access Control (MAC) sublayer that operates according
to a listen-before-talk scheme, known as Carrier Sense Multiple Access with Collision
Avoidance (CSMA/CA). A mapping of the defined IEEE 802.11 layers to the OSI reference
model is presented in Figure 2.1.
Having published its first 802.11 standard in 1997, the Working Group (WG) received
feedback that many products did not provide the degree of compatibility that customers
expected. Thus, the need for a certification program came up and led to the foundation of the
Wireless Ethernet Compatibility Alliance (WECA) in 1999, which renamed to Wi-Fi Alliance
(WFA) in 2003. Wi-Fi certification has become a well-known certification program that has a
significant market impact [8].
2.1.1
Infrared (IR)
Infrared light lies between the visible and microwave portions of the electromagnetic
spectrum. Infrared light has a range of wavelengths that lies from red light to violet. These
wavelengths correspond to a frequency range of approximately 1 to 400 THz (1 THz = 1.000
GHz). Standards regarding IR data transmission are mainly published by the Infrared Data
Association (IrDA). The infrared physical layer supports two data rates: 1 and 2 Mbit/s. The
specification of two data rates is aimed at allowing [10]:

A smooth migration to higher data rates

Asymmetric operation of the BSS
7
IEEE 802.11 Standards Family
Figure 2.1 IEEE 802.11 standards mapped to the OSI reference model [9]
The modulation method that is adopted for this physical layer is Pulse Position Modulation
(PPM) since it is very effective in wireless transmissions. For each of the two data rates
specified here, there is a different PPM scheme: 16-PPM for 1 Mbit/s and 4-PPM for 2 Mbit/s.
The purpose of this feature is to ensure that the basic pulse is the same at both data rates,
which minimizes the additional complexity introduced by the 2 Mbit/s data rates [10]. The use
of IR for WLANs is today obsolete, since it has not been adopted by any IEEE 802.11
implementation.
2.1.2
Frequency Hopping Spread Spectrum (FHSS)
FHSS is a method of transmitting radio signals by rapidly switching a carrier among many
frequency channels, using a pseudo-random sequence known to both transmitter and receiver.
In the United States, the FCC requires that at least 75 discrete frequencies must be employed
for each transmission channel and that a signal cannot remain on any particular frequency for
more than 400 ms. In the IEEE 802.11, the maximum length of a packet is around 30 ms, and
the hops are 1 MHz apart from one another. FHSS can be employed for both analog and
digital communications, but is currently implemented primarily for digital transmissions. If 75
contiguous frequencies are used, then the bandwidth required for a transmission is 75 times
larger than when only one frequency is used – the spectrum is spread over a larger portion of
the transmission band (hence “frequency hopping spread spectrum”). The original motivation
for developing
this
technique
was
a
desire
to avoid
hostile
jamming
of a
radio signal. If a transmission hops to a jammed frequency, the data sent in vain on that
8
IEEE 802.11 Standards Family
frequency are resent after the next hop. For wireless networks, FHSS has a desirable “side
effect:” It minimizes the chances that different transmitters on the network will encounter
interference from one another; otherwise the network could potentially disable itself.
The IEEE 802.11 FHSS PHY was meant to operate in the 2.4-GHz band at speeds of 1 or 2
Mbit/s. The FHSS system was discarded by the IEEE 802.11b standard after it was found that
having two transmission techniques for one standard meant that two kinds of (incompatible)
equipment were necessary to implement the standard, and DSSS turned out to be the more
reliable technique [11].
2.1.3
Direct Sequence Spread Spectrum (DSSS)
DSSS is one of the most successful data transmission techniques. Beside wireless LANs, it is,
also, used in cellular networks (CDMA systems) and Global Positioning Systems (GPS). The
idea is to multiply the data being transmitted to a pseudo random binary sequence of a higher
bit rate.
DSSS systems spread the signal energy across a relatively wide band by increasing the
occupied bandwidth. A DSSS transmitter converts a bit stream into symbol stream where each
symbol represents a number of bits depending on the Phase-Shift Keying (PSK) modulation
technique. The symbol information is converted into a complex-valued signal, which is fed to
the spreader. This spreader multiplies its input signal with a Pseudo Noise (PN) sequence,
which is called a chip sequence. The result of this multiplication is a signal with a wider
bandwidth. The in-phase and quadrature components of the spreader output signal are fed to a
quadrature modulator. The transmitter front-end provides filtering, upmixing and power
amplification. The IEEE 802.11 DSSS is based on 11-chips Barker sequence. The 11-chip
spreading makes the occupied bandwidth larger and increases the effective bandwidth from 1
MHz to 11 MHz. The IEEE 802.11 standard specifies the following bit rates: 1 Mbit/s with
BPSK, 2 Mbit/s with QPSK, 5.5 Mbit/s and 11 Mbit/s with Complementary Code Keying
(CCK) [12]. Some of the benefits of DSSS include:

Resistance to intended and unintended jamming.

Sharing of a single channel among multiple users.

Reduced signal/background-noise level hampers interception.

Determination of relative timing between transmitter and receiver.
9
IEEE 802.11 Standards Family
2.1.3.1 Complementary Code Keying (CCK)
CCK is a modulation technique that allows for multi-channel operation in the 2.4 GHz band
by virtue of using the existing IEEE 802.11 1 Mbit/s and 2 Mbit/s DSSS channelization
scheme. In 1999, CCK was adopted to supplement the Baker code in wireless digital networks
to achieve data rate higher than 2 Mbit/s at the expense of shorter distance. This is due to the
shorter chipping sequence in CCK (8 bits versus 11 bits in Barker code) that means less
spreading to obtain higher data rate but more susceptible to narrow-band interference resulting
in shorter radio transmission range. Beside shorter chipping sequence, CCK also has more
chipping sequences to encode more bits (4 chipping sequences at 5.5 Mbit/s and 64 chipping
sequences at 11 Mbit/s) increasing the data rate even further.
2.1.4
Orthogonal Frequency-Division Multiplexing (OFDM)
OFDM is a modulation technique for transmitting large amounts of digital data over a radio
wave. It extends the concept of single carrier modulation by using multiple sub-carriers within
the same single channel. The total data rate to be sent in the channel is divided between the
various sub-carriers. The data do not have to be divided evenly nor do they have to originate
from
the
same
information
source.
Advantages
include
using
separate
modulation/demodulation customized to a particular type of data, or sending out banks of
dissimilar data that can be best sent using multiple, and possibly different, modulation
schemes.
OFDM offers an advantage over single-carrier modulation in terms of narrow-band frequency
interference. Since this interference will only affect one of the frequency sub-bands, the other
sub-carriers will not be affected by the interference. Since each sub-carrier has a lower
information rate, the data symbol periods in a digital system will be longer, adding some
additional immunity to impulse noise and reflections. OFDM systems usually require a guard
band between modulated sub-carriers to prevent the spectrum of one sub-carrier from
interfering with another. These guard bands, however, lower the system’s effective
information rate when compared to a single carrier system with similar modulation [13].
OFDM was not supported by the IEEE 802.11 standard until its first amendment, the IEEE
802.11a. Nowadays, OFDM and OFDM-based modulation techniques are used by IEEE
802.11a/g/n, in addition to other 802 family standards, such as IEEE 802.16.
10
IEEE 802.11 Standards Family
2.2 IEEE 802.11 Amendments
The huge success in the market of the IEEE 802.11 standard, led to its continuous
improvement and various revisions of the original standard, driven by a complete alphabet. To
avoid confusion with other 802 standards, letters l, o, q and x are not used [8].
2.2.1
IEEE 802.11a
IEEE 802.11a is the first amendment to the original IEEE 802.11 specification, started in
September 1997 and finally approved two years later, in September 1999. It added an OFDM
physical layer that supports a higher data rate of up to 54 Mbit/s using the 5 GHz band. This
high frequency means that IEEE 802.11a signals can achieve a shorter range that the 2.4 GHz
signals and, also, they have more difficulties in penetrating walls and other obstructions.
On the other hand, using the 5 GHz band gives IEEE 802.11a a significant advantage, since
the 2.4 GHz band is heavily used to the point of being crowded, thus it is prone to conflicts
that can cause frequent dropped connections and degradation of service. Even though the
IEEE 802.11a standard has a maximum raw data rate of 54 Mbit/s, it yields realistic net
achievable throughput in the mid-20 Mbit/s. Finally, the data rate can be reduced to 48, 36,
24, 18, 12, 9 and the 6 Mbit/s (also known as Adaptive Rate Selection), if signal quality
becomes an issue.
2.2.2
IEEE 802.11b
IEEE 802.11b was developed at the same time with IEEE 802.11a. Since IEEE 802.11b
gained in popularity much faster that IEEE 802.11a did and due to its lower cost, it serves
better the home market, whereas IEEE 802.11a is usually found on business networks. IEEE
802.11b supports a bandwidth of up to 11 Mbit/s using a Direct-Sequence Spread Spectrum
(DSSS) PHY. It also uses the same CSMA/CA medium control access method defined in the
original standard. However, due to the CSMA/CA protocol overhead, in practice the
maximum IEEE 802.11b throughput that an application can achieve is about 5.9 Mbit/s using
TCP and 7.1 Mbit/s using UDP. IEEE 802.11b uses the 2.4 GHz radio band, same as the
original IEEE 802.11 standard. Using this frequency offers lower production costs, but it is
more possible to interfere with other devices that operate at the same band, such as microwave
ovens, cordless phones etc.
11
IEEE 802.11 Standards Family
2.2.3
IEEE 802.11d
With the publishing of IEEE 802.11b, the IEEE 802.11 Working Group was beginning to
realize the acceptance of its products. However, IEEE 802.11b was defined to operate only in
the United States, Canada, Japan, France, Spain, and the portion of Europe operating under
ETSI regulations. These six locations were defined in the standard as “regulatory domains”.
Outside any of these regulatory domains, a device could not be called compliant with the
IEEE 802.11 standard.
To address these regional requirements, the IEEE 802.11 Task Group d (TGd) was formed in
June 1999 and the standard was finally released two years later. The solution chosen by TGd
was to add a management protocol that would announce certain regulatory and location
information. This protocol allows mobile devices to determine whether they are allowed to
operate in that location and, if operation was allowed, to configure their radio to comply with
the local regulations [14].
2.2.4
IEEE 802.11e
Initially the goal of the IEEE 802.11e project, as approved at the end of March 2000, was to
add several enhancements to the IEEE 802.11 standard that would include efficiency
improvements, support for Quality of Service, and a higher security level. However, soon
enough, the IEEE 802.11 frame encryption algorithm WEP was broken by an attack. The
security enhancements were then displaced to a new TG, the TGi. Finally, in 2005, the IEEE
802.11e was approved targeting to provide a set of QoS enhancements through modifications
to the MAC layer [8].
In particular, the IEEE 802.11e standard introduced a new coordination function, the Hybrid
Coordination Function (HCF), in order to enhance the DCF and PCF access methods and
support applications with QoS requirements. Within the HCF, there are two methods of
channel access, similar to those defined in the legacy IEEE 802.11 MAC; the HCCA and the
EDCA. HCF is further explained in sub-chapter 3.1.3.
IEEE 802.11e also enhanced the MAC layer to support block ACK. The main idea behind this
mechanism is that the receiver instead of sending a single ACK for each data frame that has
correctly been received, it sends a block ACK after having received multiple frames. This
block ACK indicates that the last n frames were received correctly. There are two variations
of the block ACK mechanism, the immediate and the delayed block ACK.
12
IEEE 802.11 Standards Family
After sending a block of frames, the sender transmits a Block ACK Request (BAR) frame, to
which the receiver responds with a Block ACK (BA). The difference between immediate and
delayed block ACK is that in the first case the BAR solicits an immediate BA response, while
with the delayed block ACK the BAR itself needs to be acknowledged, and then the BA is
transmitted which in turn needs to be acknowledged [15]. Figure 2.2 demonstrates the
operation of both mechanisms.
Figure 2.2 Immediate and delayed block ACK sessions [15]
2.2.5
IEEE 802.11f
In 2000, IEEE 802.11 Task Group f (TGf) started to work on a recommended practice for a
protocol to be used between IEEE 802.11 APs, the Inter Access Point Protocol (IAPP), today
known as IEEE 802.11.1. The goals of IEEE 802.11f are to provide a recommended way for
APs to communicate with each other when a mobile station (STA) roams between them, to
describe a way for an AP to update the forwarding tables in IEEE 802.1 MAC bridges that
may be included in the IEEE 802.11 DS, to establish a format for APs to exchange context
information about a station that has roamed, and to enable the distribution of station context
information from one AP to neighboring APs. Some of the things that IEEE 802.11f does not
define is what an AP should put into the context container in order to communicate with a
13
IEEE 802.11 Standards Family
peer AP or which format should have the information in the context container. This was
purposely left for future standardization [14]. IEEE 802.11f was a Trial Use Recommended
Practice and the IEEE 802 Executive Committee approved its withdrawal on February 2006.
2.2.6
IEEE 802.11g
IEEE 802.11g is the third modulation standard for wireless LANs and was finally approved on
June 2003. Since IEEE 802.11a is operating in the 5 GHz band, it lacks of backward
compatibility with plain IEEE 802.11 devices. This is one of the main reasons that led to the
formation of IEEE 802.11g. Basically, IEEE 802.11g attempts to combine the advantages of
both 802.11a and 802.11b. It uses the 2.4 GHz band, just like IEEE 802.11b does, but it
reaches a maximum raw data rate of 54 Mbit/s using the OFDM modulation scheme, as IEEE
802.11a does.
802.11g received major acceptance and was rapidly adopted by customers even before its
ratification, due to the need for higher speeds and reduced manufacturing costs. However, it
still suffers from the same interference as IEEE 802.11b, in the already crowded 2.4 GHz
range.
2.2.7
IEEE 802.11h
In 1999, the European Union modified its regulations that applied to the radio band used by
IEEE 802.11a. The new regulations required that any wireless device or system operating in
the 5 GHz band must implement four (4) new functions, in order to protect existing civil and
military radar that already operates in the band, as well as to minimize the “hotspot” that
might show up in urban areas in radar images of earth satellite systems and also use this band.
These four requirements are:
i. The device must be able to detect the presence of radar operations.
ii. The device or system must be able to avoid interfering with radar operations.
iii. The system must be able to uniformly spread its operation across all the channels that
may be used in the band.
iv. The system must be able to minimize the overall power output of the system.
To ensure that IEEE 802.11 WLANs operated in a consistent fashion when implementing
systems to the new regulations, the IEEE 802.11 Task Group h (TGh) was chartered to write
an amendment to the IEEE 802.11 standard. The amendment was ratified by the IEEE in July
2003 [14].
14
IEEE 802.11 Standards Family
2.2.8
IEEE 802.11i
Late in 1999, several weaknesses of the Wired Equivalent Privacy (WEP) algorithm of the
initial IEEE 802.11 standard began to arise. Additionally, its pre-shared keying concept did
not allow for integration into enterprise networks, where each device should have its own
unique key. Thus, companies required their employees to use Virtual Private Networks
(VPNs) and IEEE 802.11 became synonym for insecurity [8].
As mentioned earlier, in 2000 the TGe was approved, aiming to work on both security and
QoS enhancements. In May 2001, TGi was split from TGe to focus only on security
enhancements. The task for TGi was two-fold [14]:

To create a new, very secure means of authentication and privacy that would no be
vulnerable to the weaknesses of WEP.

To enhance the security characteristics of the existing WEP.
Meanwhile, the Wi-Fi Alliance (WFA) did not wait for a solution and started its Wi-Fi
Protected Access (WPA) certification program. WPA was an intermediate solution to WEP
insecurities, that implemented a subset of a draft of IEEE 802.11i. IEEE 802.11i was finally
approved in June 2004 and the result of its work were the Robust Security Network (RSN)
and the Transition Security Network (TSN).
The WFA refers to their approved, interoperable implementation of the RSN, as WPA2. RSN
has been developed to provide very strong encryption and strong per-packet authenticity. It
also includes an encryption scheme designed from scratch, that relies on the Advanced
Encryption Standard (AES), whereas both WEP and WAP make use of the RC4 stream cipher
mechanisms. However, it should to be noted that IEEE 802.11i does not address all of the
security problems that plague a WLAN. IEEE 802.11i specifically addresses authentication
and confidentiality of data frames but it does not address other shortcomings of IEEE 802.11,
such as protection of management frames, prevention of denial of service attacks, or
prevention of attacks that may take place in higher layers, e.g., ARP spoofing [14].
Old hardware cannot be upgraded to make use of RSN. Thus, legacy devices would not be
able to coexist in the same network with newer ones that make use of RSN. TSN provides a
method for legacy equipment that is capable only of WEP encryption to operate together with
equipment that provides RSN capabilities, in a mixed environment.
15
IEEE 802.11 Standards Family
2.2.9
IEEE 802.11j
In August 2002, the Japanese government expanded the frequency band of operation beyond
the 5.15-5.25 GHz range, to include both the licensed and unlicensed spectrum from 4.9 GHz
to 5.25 GHz, for operating WLANs inside the country. Approved in December 2002 and
ratified in September 2004, IEEE 802.11j describes the necessary means to comply with the
new Japanese regulatory requirements. and its new applications. The specifications for IEEE
802.11j were defined so that legacy IEEE 802.11a 5 GHz WLANs would require only
minimal changes to the PHY hardware and software of the radios [14].
2.2.10
IEEE 802.11k
Approved in December 2002 and ratified in March 2008, IEEE 802.11k is an amendment for
radio resource management. It is intended to improve the way traffic is distributed within a
network, by defining a series of measurement requests and reports that can be used in layers
above the MAC. IEEE 802.11k enables a radio network to collect information regarding other
APs and link quality to neighbor stations. It also provides methods to measure interference
levels and medium load statistics [16]. In a wireless LAN if there are more than one APs, a
station will connect to the one with the strongest signal. However, this can sometimes lead to
excessive demand on this AP, while the others remain underutilized. In a network conforming
to IEEE 802.11k, if the AP having the strongest signal is loaded to its full capacity, a
connected station can switch to one the underutilized APs. Even though the signal may be
weaker, the overall throughout is greater because of the more efficient use of network
resources. The steps for a station to switch to a new AP are the following:

The associated AP determines that the station should switch to another AP.

The AP informs the station to prepare to switch to a new AP.

The STA requests a list of the nearby APs.

The AP provides the STA with the requested list.

The station moves to the best AP based on the reported list.
2.2.11
IEEE 802.11n
IEEE 802.11n amendment defines modifications to both IEEE 802.11 PHY and MAC layer,
aiming to a significant increase of the maximum data rate. In particular, IEEE 802.11n can
derive a throughput from 100 Mbit/s up to 600 Mbit/s, which is far beyond the maximum
throughput of 54 Mbit/s that was supported by then. To achieve that, IEEE 802.11n inherited
16
IEEE 802.11 Standards Family
some new technologies to the existing IEEE 802.11 standard, with the most notable being the
so called Multiple-Input Multiple-Output (MIMO). MIMO is a form of “smart antenna”
technology. It can use up to four antennas, to resolve more data than possible using a single
antenna (Figure 2.3). To provide this, it uses either Spatial Division Multiplexing (SDM), or
beam forming.
Figure 2.3 MIMO with TXN transmit and RXN receive antennas [14]
MIMO systems offer a number of benefits to WLANs, such as [14]:

Using several antennas, the signal strength can be significantly improved.

The interference probability is decreased.

Higher channel capacity.

The receiver has the possibility to recover lost or corrupted data using different signal
paths.
Another feature of IEEE 802.11n is the usage of optional 40 MHz channels, instead of only 20
MHz channel widths, as it was defined in the previous IEEE 802.11 PHYs. The 40 MHz
operation raised many concerns regarding neighbor friendly behavior. Especially for the
crowded 2.4 GHz mode, there were concerns that it would severely affect the performance of
existing IEEE 802.11 and non-802.11 devices, such as Bluetooth (802.15.1) and ZigBee
(802.15.4) [8]. As a consequence, the development of an appropriate compromise was a
prerequisite for IEEE 802.11n's ratification, which was finally made six years later, in
September 2009. As a result, IEEE 802.11n devices can operate in 40 MHz mode only if they
are able to detect 20 MHz-only devices. They can also be enabled to operate in the 5 GHz
band or, alternatively, in the 2.4 GHz only if there is knowledge that they will not interfere
with other devices.
17
IEEE 802.11 Standards Family
Finally, IEEE 802.11n introduces enhancements in the MAC layer, too, with the main
mechanisms being frame aggregation and block acknowledgment. With frame aggregation
multiple data packets from the Network layer (OSI layer 3) are combined in one larger frame
for transmission. Frame aggregation achieves better channel utilization from the legacy frame
encapsulation mechanism, since there is reduced header overhead, in addition to inter-frame
time. IEEE 802.11n supports a two-level aggregation technique. At the top of the MAC there
is the MAC Protocol Service Unit (MSDU) aggregation (A-MSDU), which forms a MAC
Protocol Data Unit (MPDU). Then, in the second level, at the bottom of the MAC there is the
MPDU aggregation (A-MPDU) which forms a PSDU that is passed down to the PHY layer
[15].
As mentioned earlier, in sub-chapter 2.2.4, block acknowledgment was initially introduced by
the IEEE 802.11e. IEEE 802.11n enhances this mechanism to support frame aggregation and
further improve the network efficiency. The resulted mechanisms are referred to as HTimmediate block ACK and HT-delayed block ACK, where HT stands for High Throughput.
2.2.12
IEEE 802.11p
Published in 2010, IEEE 802.11p is an approved amendment to the IEEE 802.11 standard to
add Wireless Access in Vehicular Environments (WAVE). It describes several modifications
brought to the PHY and MAC layers of the original standard in order to support Intelligent
Transportation Systems (ITS) applications. ITS basically refers to information being
exchanged between high speed vehicles or between vehicles and roadside infrastructure
aiming to improve transport outcomes, such as safety, reliability and comfort. IEEE 802.11p
operates in 5.855-5.925 GHz band in the United States, and the 5.855-5.905 GHz band in
Europe, which was recently harmonized for IEEE 802.11p operation. Its PHY is similar to
OFDM-based IEEE 802.11a, with the major difference that IEEE 802.11p operates on 10
MHz wide channels, instead of 20 MHz.
2.2.13
IEEE 802.11r
With the advancements of Wireless LANs, fast roaming support has become one of the most
important issues in IEEE 802.11. The TGr was formed in 2004 to address roaming capabilities
of real-time applications with an attempt to minimize BSS transition time, while still
providing the IEEE 802.11i security and IEEE 802.11e QoS. IEEE 802.11r permits seamless
connectivity with a fast and secure handover from one AP to another, within the same
18
IEEE 802.11 Standards Family
mobility domain [17]. With IEEE 802.11r, a station can negotiate security and QoS settings
with neighbor APs, while it is still associated to another AP. Thus, if a station loses
connectivity with the AP and re-associates to a neighbor AP, the duration of the reconnectivity can be substantially reduced. IEEE 802.11r was finally approved in June 2008.
2.2.14
IEEE 802.11s
IEEE 802.11s is an amendment to the IEEE 802.11 standard for mesh networking. The project
got approved in 2004 and in June 2011, the TGs Draft 12.0 closed with 97.2% approval rate.
IEEE 802.11s defines a multi-hop framework, where wireless devices can interconnect to
create a WLAN mesh network, by serving as wireless routers. Precisely, it defines
mechanisms that:

Determine how to route packets through the mesh.

Allow access to external networks, such as the Internet.

Allow stations to access the mesh network.
2.2.15
IEEE 802.11w
IEEE 802.11w is the most recent security related amendment and was released in September
2009. It consists an amend to the IEEE 802.11i, and its objective is to protect against subtle
attacks on wireless LAN management frames. For example, it protects against network
disruption caused by malicious systems that forge disassociation requests, that appear to be
sent by valid equipment. This is achieved by providing data confidentiality of management
frames, mechanisms that enable data integrity, data origin authenticity and replay protection.
2.2.16
IEEE 802.11y
In 2005 the FCC established a new regulatory regime for systems that operate within the 3.653.70 GHz band [8]. IEEE 802.11y, which approved for publication in June 2008, enables
high-powered Wi-Fi equipment to operate in the 3.65-3.7 GHz band in the United States.
With up to 20 W output power, stations operating at the US 3.65 GHz band can communicate
at distances of 5 km or more. IEEE 802.11y adds three new concepts to the IEEE 802.11 base
standard:

Contention Based Protocol (CBP): enhancements to carrier sensing and energy
detection mechanisms.
19
IEEE 802.11 Standards Family

Extended Channel Switch Announcement (ECSA): a mechanism for an AP to notify
the associated stations of its intention to change channels or to change channel
bandwidth.

Dependent Station Enablement (DSE): is the process by which an enabling station
grants permission and dictates operational procedures to dependent stations.
Title
Project
Final
Title
Comment
approval date approval date
802.11a 16-09-1997
16-09-1999
Higher Speed PHY
54 Mbit/s OFDM PHY in 5
Extension in the 5GHz Band GHz
802.11b 09-12-1997
16-09-1999
Higher Speed PHY
11 Mbit/s DSSS PHY in 2.4
Extension in the 2.4 GHz
GHz
Band
802.11d 26-06-1999
14-06-2001
Operation in Additional
Allows devices to comply
Regulatory Domains
with regional requirements
802.11e 30-03-2000
22-09-2005
MAC Enhancements QoS
Support for QoS
802.11f 30-03-2000
12-06-2003
Inter-Access Point Protocol
Released as 802.11.1 and
Across Distribution Systems administratively withdrawn
Supporting IEEE 802.11
on 03-02-2006
Operation
802.11g 21-09-2000
12-06-2003
Further Higher Data Rate
54 Mbit/s OFDM PHY in
Extension in the 2.4 GHz
2.4 GHz
Band
802.11h 07-12-2000
11-09-2003
Spectrum and Transmit
In Europe, 5 GHz devices
Power Management
must implement 802.11h
Extensions in the 5 GHz
Band in Europe
802.11i 30-05-2001
24-06-2004
MAC Security
MAC security
Enhancements
enhancements, known as
WPA and WPA2
802.11j 11-12-2002
802.11k 11-12-2002
802.11n 11-09-2003
23-09-2004
31-03-2008
11-09-2009
4.9 GHz-5 GHz Operation in Compliance with Japanese 5
Japan
GHz spectrum regulation
Radio Resource
Measurements of the
Measurement
wireless channel
High Throughout
600 Mbit/s MIMO PHY in
2.4 GHz and 5 GHz
20
IEEE 802.11 Standards Family
802.11p 23-09-2004
802.11r 13-05-2004
15-07-2010
30-06-2008
Wireless Access for the
Car to car communication,
Vehicular Environment
closely related to IEEE 1609
Fast Roaming
Fast hand-off for moving
devices
802.11s 13-05-2004
01-08-2011
Mesh Networking
Transparent multi-hop
operation
802.11w 20-03-2005
802.11y 16-03-2006
30-09-2009
30-06-2008
Protected Management
Security for management
Frames
frames
3650-3700 MHz Operation
Contention-based protocols
in USA
for FCC band 3.65 GHz in
the U.S.
Table 2.1 Published amendments [8]
2.3 Ongoing standardizations
2.3.1
IEEE802.11ac
IEEE 802.11ac is an emerging amendment that aims to push WLAN throughput over the
gigabit-per-second barrier and up to 7 Gb/s. The following new technologies are introduced in
the IEEE 802.11ac draft standard, among others:

Multi-user multiple-input multiple-output (MU-MIMO)

Larger channel bandwidths of 80 and 160 MHz
Most wireless networks have multiple active clients that share the available bandwidth. If this
sharing is done in time, then the overall throughput can only be increased by increasing the
link rate for all clients. However, many clients may not be able to transmit at the highest IEEE
802.11ac rates. For such clients, MU-MIMO is the solution to get significant network
throughput gains. A MU-MIMO capable transmitter can transmit multiple packets
simultaneously to multiple clients. In IEEE 802.11ac, a MU-MIMO mode is defined with up
to eight spatial streams divided across up to four different clients. Thus, the data rate per client
can be up to four times larger, because the MU-MIMO packets can be transmitted at the
maximum data rate per client while without MU-MIMO, each client can only be transmitted
to about a quarter of the time such that the effective per-user throughput is a quarter of its
maximum [18].
IEEE 802.11ac will operate in 80 MHz and 160 MHz channel bandwidths, in addition to the
20 and 40 MHz defined in IEEE 802.11n. IEEE 802.11ac, 2 more bandwidths of 80 and 160
21
IEEE 802.11 Standards Family
MHz are introduced. The 80 MHz mode uses two adjacent 40 MHz channels with some extra
subcarriers to fill the unused tones between two adjacent 40 MHz channels. The 160 MHz
mode uses two separate 80 MHz channels without any tone filling in the middle of these two
sub-channels. The two 80 MHz channels do not have to be adjacent. This increases the
probability of finding a 160 MHz channel at the cost of additional hardware to send and
receive in two non-adjacent 80 MHz channels. IEEE 802.11ac only applies to the 5 GHz
band, since there is no room in the 2.4 GHz band for the 80 and 160 MHz channels [18]. IEEE
802.11ac also provides backward compatibility with IEEE 802.11a and IEEE 802.11n
devices.
2.3.2
IEEE 802.11ad
The objective of IEEE 802.11ad is to define standardized modifications to both IEEE 802.11
PHY and MAC layers, in order to enable operation in the 60 GHz frequency band, capable of
very high throughput in short-range data transmissions. 60 GHz band can provide solutions
for various applications, such as local file transfer and HD video transfer. The project got
approved in December 2008 and is expected to be published by the end of 2012.
2.3.3
IEEE 802.11ae
IEEE 802.11ae is the second amendment for QoS enhancements to the IEEE 802.11 base
standard. The project was approved in December 2009 and was finally approved in the first
half of 2012. Its goal is to provide QoS management and prioritization of management frames.
2.3.4
IEEE 802.11af
IEEE 802.11af defines mechanisms for operation of WLAN within the TV white space, that
is, the TV unused spectrum. The work started in December 2009 and by July 2011, the Draft
1.06 met with 62% approval rate. Some of the advantages of operating the TV white space
are:

Propagation characteristics: In view of the fact that the TV white spaces would use
frequencies below 1 GHz, this would allow for greater distances to be achieved.

Additional bandwidth: One of the advantages of using TV white space is that unused
frequencies can be accessed. However, it will be necessary to aggregate several TV
channels to provide the bandwidths that Wi-Fi uses on 2.4 and 5 GHz, to achieve the
required data throughput rates.
22
IEEE 802.11 Standards Family
2.3.5
IEEE 802.11ai
IEEE 802.11ai is an amendment approved in December 2010 for fast initial link setup. The
project intends to amend the MAC only, thereby providing a fast initial link service to all
IEEE 802.11 devices. One way to achieve that, is by reducing the number of messages
transmitted per user during set-up, to free airtime and increase the number of users that may
simultaneously enter an ESS. The project does not intend to remove any existing mechanisms
of IEEE 802.11 but to add additional, coexisting functionality, enabling fast initial link set-up.
The project is still in a very early phase and no draft has been released, yet.
2.3.6
IEEE 802.11ah
This amendment defines an OFDM physical layer operating in the license-exempt bands
below 1 GHz, excluding the TV white space band. The data rates defined in this amendment
optimize the rate vs range performance of the specific channelization in a given band. IEEE
802.11ah adds support for transmission range up to 1 km and data rate higher that 100 kb/s.
As IEEE 802.11ai, IEEE 802.11ah is, also, in a very early phase and no draft has been
released, yet.
Title
Project
Expected final
Title
Comment
approval date approval date
802.11ac 26-09-2008
31-12-2012
Very High
Enhancements for > 1 Gb/s
Throughout 6 GHz
throughput for operation in bands
below 6 GHz
802.11ad 10-12-2008
31-12-2012
Very High
Enhancements for > 1 Gb/s
Throughout 60 GHz throughput for operation in 60
GHz band
802.11ae 09-12-2009
31-12-2012
Prioritization of
QoS management and
Management Frames prioritization of management
frames
802.11af 09-12-2009
31-12-2013
TV White Spaces
WLAN operation within the TV
white spaces
802.11ai 08-12-2010
802.11ah 04-10-2010
13-12-2014
31-12-2013
Fast Initial Link
Provide a fast initial link service
Setup
to all IEEE 802.11 devices
Sub 1 GHz
PHY operation in the bands below
1 GHz
Table 2.2 Amendments in process
23
Chapter 3
IEEE 802.11 Medium Access Control (MAC) Layer
The Medium Access Control (MAC) is a sublayer of the data link layer specified in the sevenlayer OSI model (layer 2). Among other functions, it provides addressing and channel access
control mechanisms that make it possible for several stations to communicate on multi-access
networks. Similar to the Ethernet MAC, the IEEE 802.11 MAC uses a Carrier Sense Multiple
Access (CSMA) scheme to control access to the transmission medium. However, collisions
waste valuable transmission capacity, so instead of the collision detection (CSMA/CD)
employed by Ethernet, IEEE 802.11 uses Collision Avoidance (CA). Also like Ethernet, IEEE
802.11 uses a distributed access scheme with no centralized controller. That is, each IEEE
802.11 station uses the same method to gain access to the medium [19].
The architecture of the MAC sublayer includes the Distributed Coordination Function (DCF),
the Point Coordination Function (PCF), the Hybrid Coordination Function (HCF) and their
coexistence in an IEEE 802.11 LAN (Figure 3.1). These functions are further explained in the
following sub-chapters.
Figure 3.1 MAC architecture [9]
24
IEEE 802.11 Medium Access Control (MAC) Layer
3.1 Distributed Coordination Function (DCF)
The basic medium access protocol is a DCF that allows for automatic medium sharing
between compatible PHYs through the use of CSMA/CA and a random backoff time
following a busy medium condition. In addition, all individually addressed traffic uses
immediate positive acknowledgement (ACK frame) where retransmission is scheduled by the
sender if no ACK is received [9].
Figure 3.2 Distributed Coordination Function [9]
3.1.1 Carrier Sense Multiple Access (CSMA)
CSMA is a probabilistic protocol in which a station listens to the medium before transmitting.
If the medium is sensed as “idle”, the station may proceed to the transmission. Otherwise, the
medium is sensed as “busy” and the station postpones its transmission. IEEE 802.11 defines
two carrier sensing mechanisms to avoid interference in wireless local area networks for the
kind of interference originating from within the receiving range of a receiver [21]. The two
mechanisms and the function each one uses are summarized in Table 3.1.
Carrier Sensing
Function
Description
Mechanism
Physical carrier sensing
Virtual carrier sensing
Clear Channel Assessment
Indicates a busy medium for the
(CCA)
current frame
Network Allocation Vector
Reserves the medium as busy for
(NAV)
future frames
Table 3.1 IEEE 802.11 Carrier Sensing Mechanisms
The CCA is composed of two related functions; Carrier Sense (CS) and Energy Detection
(ED). The first one refers to the ability of a station to detect and decode incoming Wi-Fi
25
IEEE 802.11 Medium Access Control (MAC) Layer
signal preamble, while ED refers to the ability of a station to detect non-Wi-Fi energy level
present on the current channel. Such an energy level may be produced by external sources that
interfere. Unlike CS which can determine the exact length of the time the medium will be
busy with the current frame, ED must sample the medium every slot time to determine if the
energy still exists [21]. While CCA itself is implemented at the PHY layer, the primary impact
of its performance and complexity is on MAC metrics like throughput and energy efficiency.
The channel status is determined by the sensed signal power level in the channel. If the power
level is above a predefined threshold, the medium is considered to be busy, otherwise idle.
In addition to CCA determining the medium idle/busy state for the current frame and noise,
the NAV allows stations to indicate the amount of time required for transmission of required
frames immediately following the current frame. It is important to reserve the medium as busy
for these mandatory frames. The importance of NAV virtual carrier sense is to ensure medium
reservation for frames critical to the operation of the IEEE 802.11 protocol. Typically these
are control frames, but not always. They include IEEE 802.11 acknowledgments, subsequent
data and acknowledgment frames as part of a fragment burst, and data and acknowledgment
frames following an RTS/CTS exchange [21]. Finally, the medium is determined to be idle
only when both the physical and virtual carrier sense mechanisms indicate it to be so [15].
3.1.2 Collision Avoidance (CA), Random Backoff & Contention Window
The purpose of CA is to improve the performance of CSMA by attempting to divide the
wireless channel somewhat equally among all transmitting nodes within the collision domain.
Every station that wishes to transmit first senses the medium for a fixed duration, called DCF
Inter-Frame Space (DIFS). If the medium is idle the station assumes it may take ownership of
the medium and begin a frame exchange sequence. If the medium is busy, the stations waits
for the medium to go idle, defers for DIFS, and waits for a further random backoff period. If
the medium remains idle for the DIFS deferral and the backoff period, the stations assumes
that it may take ownership of the medium and begin a frame exchange sequence [21].
The random backoff period provides the CA aspect. When the network is loaded, multiple
stations may be waiting for the medium to go idle having accumulated packets to send while
the medium was busy. Since each station probabilistically selects a different backoff interval,
collisions where more than one stations begin transmission at the same time are unlikely. The
length of the backoff period is determined by the following equation:
26
IEEE 802.11 Medium Access Control (MAC) Layer
BackoffTime = Random() x aSlotTime
where:
Random() = Pseudo-random integer drawn from a uniform distribution over the interval [0,
CW], where CW (Contention Window) is an integer within the range of values
of the PHY characteristics aCWmin and aCWmax, aCWmin ≤ CW ≤ aCWmax.
It is important that designers recognize the need for statistical independence
among the random number streams among stations.
aSlotTime = The value of the correspondingly named PHY characteristic.
Once CW reaches aCWmax, it shall remain at the value of aCWmax until it (CW) is reset.
This improves the stability of the access protocol under high-load conditions [9].
Figure 3.3 CW incrementation example [9]
27
IEEE 802.11 Medium Access Control (MAC) Layer
Figure 3.4 Backoff Procedure for DCF [9]
3.1.3 Request to Send/Clear to Send (RTS/CTS) Frame Exchange
CA cannot detect transmissions from hidden nodes. RTS/CTS, on the other hand, helps to
solve this problem (the hidden node problem). To protect its transmission a station may begin
a sequence with an RTS/CTS exchange. During this sequence a station wishing to transmit
sends an RTS frame after it has sensed the medium as idle. The destination station responds to
the RTS frame with a CTS frame after a SIFS (Short Interframe Spacing – used for highpriority transmissions) period has elapsed. Upon successful reception of the CTS frame, the
transmitting station assumes that the medium is reserved and may start sending the actual
data. Any other station that receives an RTS or CTS frame reads the duration field and sets its
NAV accordingly, so that it refrains from sending data for the given time. Also, RTS and CTS
are short frames which occupy less air time than the data frames and are, thus, less susceptible
to collisions.
Figure 3.5 RTS/CTS/data/ACK and NAV setting [9]
28
IEEE 802.11 Medium Access Control (MAC) Layer
Figure 3.6 Simplified algorithm of CSMA/CA
3.2 Point Coordination Function (PCF)
The IEEE 802.11 MAC may also incorporate an optional access method called PCF, which is
only employed for infrastructure network configurations. This access method uses a Point
Coordinator (PC), which shall operate at the AP of the BSS, to determine which station
currently has the right to transmit. The operation utilizes polling, with the PC performing the
role of the polling master.
The PCF uses a virtual CS mechanism aided by an access priority mechanism. The PCF shall
distribute information within Beacon management frames to gain control of the medium by
setting the NAV in stations. In addition, all frame transmissions under the PCF may use an
interframe space (IFS) that is smaller than the DIFS for frames transmitted via the DCF. The
use of a smaller IFS implies that the point-coordinated traffic shall have priority access to the
medium over stations in overlapping BSSs operating under the DCF access method [9].
29
IEEE 802.11 Medium Access Control (MAC) Layer
Figure 3.7 Point Coordination Function [9]
3.3 Hybrid Coordination Function (HCF)
The IEEE 802.11e is an approved amendment to the IEEE 802.11 standard that defines a set
of Quality of Service (QoS) enhancements for the wireless LAN applications through
modifications to the MAC layer. QoS is the ability to provide different priority to different
applications, users, or data flows, or to guarantee a certain level of performance to a data flow.
The IEEE 802.11e extends the DCF and PCF, through a new coordination function, namely
HCF, that enhances QoS functionality to QoS aware applications. HCF is only usable in QoS
network configurations and shall be implemented in all QoS stations. The HCF combines
functions from the DCF and PCF with some enhanced, QoS-specific mechanisms and frame
subtypes to allow a uniform set of frame exchange sequences to be used for QoS data
transfers during both the Contention Period (CP) and the Contention Free Period (CFP). The
HCF uses both a contention-based channel access method, called the Enhanced Distributed
Channel Access (EDCA) mechanism for contention-based transfer and a controlled channel
access, referred to as the HCF Controlled Channel Access (HCCA) mechanism, for
contention-free transfer [9].
3.3.1 HCF Contention Based Channel Access (EDCA)
The EDCA mechanism provides differentiated, distributed access to the wireless medium for
stations using eight different User Priorities (UPs). The EDCA mechanism defines four access
categories (ACs) that provide support for the delivery of traffic with UPs at the stations [9].
The mapping between the UPs and the ACs is demonstrated in Table 3.2.
30
IEEE 802.11 Medium Access Control (MAC) Layer
Priority
UP (Same as
802.1D
802.1D user
designation
AC
Designation
(informative)
priority)
Lowest
Highest
1
BK
AC_BK
Background
2
-
AC_BK
Background
0
BE
AC_BE
Best Effort
3
EE
AC_BE
Best Effort
4
CL
AC_VI
Video
5
VI
AC_VI
Video
6
VO
AC_VO
Voice
NC
AC_VO
Table 3.2 UP-to-AC mappings [9]
Voice
7
Every station maintains four transmit queues, one per AC as illustrated in Figure 3.8. For each
AC, an enhanced variant of the DCF, called an Enhanced Distributed Channel Access
Function (EDCAF), contends for TXOPs using a set of EDCA parameters from the EDCA
Parameter Set element received from the associated AP, or from the default values for the
parameters if no EDCA Parameter Set element is received.
31
IEEE 802.11 Medium Access Control (MAC) Layer
Figure 3.8 Reference implementation model [22]
The EDCA Parameter Set element is used by the AP to establish policy, to change policies
when accepting new stations or new traffic, or to adapt to changes in offered load. The EDCA
Parameter Set includes the following parameters [22]:

Minimal CW value for a given AC (CWmin[AC]): CWmin can be different for
different ACs. Assigning smaller values of CWmin to high priority classes can ensure
that high-priority classes obtain more TXOPs than low-priority ones.

Maximal CW value for a given AC (CWmax[AC]): Similar to CWmin, CWmax is
also on a per AC basis.

Arbitration Interframe Space (AIFS[AC]): Each AC starts its backoff procedure after
the channel is idle for a period of AIFS[AC] instead of DIFS. The AIFS[AC] for a
given AC should be equal to an SIFS plus multiple time slots (i.e., AIFS[AC] =
sSIFSTime + AIFSM[AC]*aSlotTime). Considering in legacy IEEE 802.11 we have
32
IEEE 802.11 Medium Access Control (MAC) Layer
DIFS = aSIFSTime +2*aSlotTime, AIFSN[AC] is typically set to not less than 2 such
that the shortest waiting time is DIFS.

TXOPlimit[AC]: TXOPs obtained via EDCA are referred as EDCA-TXOPs. During
an EDCA-TXOP, a station may be allowed to transmit multiple data frames from the
same AC with a SIFS gap between an ACK and the subsequent data frame
transmission. TXOPlimit[AC] gives the limit for such a consecutive transmission.

Virtual collision: If the backoff counters of two or more collocated ACs in one station
elapse at the same time, a scheduler inside the station treats the event as a virtual
collision. The TXOP is given to the AC with the highest priority among the
“colliding” ACs, and the other colliding ACs defer and try again later as if the
collision occurred in real medium.
AC
CWmin
CWmax
AIFSN
TXOP Limit
TXOP Limit
DC-CCK/PBCC
OFDM/CC-
PHY
OFDM PHY
AC_BK aCWmin
aCWmax
7
0
0
AC_BE aCWmin
aCWmax
3
0
0
AC_VI (aCWmin+1)/2 - 1 aCWmin
2
6.016 ms
3.008 ms
AC_VO (aCWmin+1)/4 - 1 (aCWmin+1)/2 -1 2
3.264 ms
1.504 ms
Table 3.3 Default EDCA Parameter Set element parameter values [9]
The QoS AP announces the EDCA Parameter Set element in all Beacon frames occurring
within two or more Delivery Traffic Indication Message (DTIM) periods following a change
in AC parameters to assure that all stations are able to receive the updated EDCA parameters.
Moreover, the EDCA Parameter Set element is included in all Probe Response and
(Re)Association Response frames. If no such element is received, the stations shall use the
default values for the parameters.
3.3.2 HCF Controlled Channel Access (HCCA)
Unlike EDCA and TXOP, HCCA is an optional enhancement for IEEE 802.11e networks. In
fact, few (if any) APs currently available are enabled for HHCA. The HCCA mechanism uses
a QoS-aware centralized coordinator, called Hybrid Coordinator (HC), and operates under
rules that are different in several significant ways from the PC of the PCF, although it may
optionally implement the functionality of a PC. In contrast to PCF, in which the interval
33
IEEE 802.11 Medium Access Control (MAC) Layer
between two beacon frames is divided into two periods of CFP and CP, the HCCA allows for
CFPs being initiated during a CP. This kind of CFP is called a Controlled Access Phase
(CAP) in IEEE 802.11e. A CAP is initiated by the AP whenever it wants to send a frame to a
station or receive a frame from a station in a contention-free manner. In fact, the CFP is a
CAP, too. The HCF protects the transmissions during each CAP using the virtual CS
mechanism. A station may initiate multiple frame exchange sequences during a polled TXOP
of sufficient duration to perform more than one such sequence. The use of virtual CS by the
HC provides improved protection of the CFP.
HCCA allows for the reservation of TXOPs with the HC. A station based on its requirements
requests the HC for TXOPs, both for its own transmissions as well as for transmissions from
the AP to itself. The HC either accepts or rejects the request based on an admission control
policy. If the request is accepted, the HC schedules TXOPs for both the AP and the station.
For transmissions from the station, the HC polls the station based on the parameters supplied
by the station at the time of its request. For transmissions to the station, the AP directly
obtains TXOPs from the collocated HC and delivers the queued frames to the station, again
based on the parameters supplied by the station.
Generally, HCCA is considered to be the most advanced and complex coordination function,
allowing QoS to be configured with great precision. QoS-enabled stations have the ability to
request specific transmission parameters, such as data rate, jitter, etc. This should allow
advanced applications like VoIP and video streaming to work more effectively on a Wi-Fi
network.
34
Chapter 4
Stream Classification Service and Interworking with 802.1AVB
4.1 Stream Classification Service
The Stream Classification Service aims to cover two of the targets within the scope of the
IEEE 802.11aa amendment:

Intra-Access Category prioritization: the need to differentiate between separate
streams within the same access category

Graceful degradation: In situations with insufficient channel capacity, there needs to
be a mechanism that marks the less important information as discardable, providing a
more graceful video degradation.
As mentioned in sub-chapter 3.1.3.1, the EDCA function in IEEE 802.11e, that provides QoS
for video and voice streams, specifies four Access Categories (AC) with different priorities.
The differentiation is enforced by configuring the contention parameters used by the station to
content for the medium, which are the minimum and maximum size of the Contention
Window (CWmin and CWmax), the Arbitration Inter-Frame Space (AIFS) and the maximum
size of the Transmission Opportunity (TXOP). Higher priority access categories have a
smaller contention window and shorter inter-frame space, increasing the probability of the
frame taking hold of the medium.
The problem of the IEEE 802.11e EDCA function in the case of video transmission is that
with only four (4) access categories, there is only one category reserved for voice traffic and
one for video streams. However, there may be a need for simultaneous transmission of many
video streams with different performance restrictions. For example, video conferencing has a
smaller tolerance for delay and jitter than streaming video. In the current EDCA scheme, these
streams should belong to the same access category and there is no way to give one stream
higher priority over the other. Also, the smaller contention window for the video access
category means that there is a higher collision probability in the presence of multiple streams
that leads to retransmissions and throughput degradation.
35
Stream Classification Service and Interworking with 802.1AVB
4.1.1
Intra-access category prioritization
The target of the IEEE 802.11aa Task Group was to increase the granularity of the EDCA
access categories without making major changes to the EDCA function of the IEEE 802.11
standard. To achieve this, it provides an alternate transmit queue for the two highest priority
access categories; voice and video. This allows individual streams with different requirements
that belong to the same access category to access the medium with different priority. Figure
4.1 shows how the EDCA function operates when intra-access category prioritization is used.
The packets for the primary and the alternate queue in each access category are kept in
separate buffers. A scheduling function selects a frame between the primary and the alternate
queue before it is passed on to the video or voice EDCA function, with a higher probability to
select a frame from the queue that has a higher User Priority. Having six queues allows for a
better mapping with the 8 user priorities specified in IEEE 802.11, which are identical to the 8
priority tags in 802.1D and therefore enable better interworking of IEEE 802.11 with other
networks (Table 4.1). The incoming frames can be classified to an alternate queue either by
specifying directly the User Priority or by using the Intra-Access Category Priority element
when the Traffic Stream is set-up (Figure 4.2).
Figure 4.1 EDCA operation with Intra-Access Category Prioritization [9]
36
Stream Classification Service and Interworking with 802.1AVB
Figure 4.2 Intra-Access Category Priority Element
Priority
Lowest
Highest
4.1.2
UP (Same as
802.1D
802.1D UP)
designation
1
BK
2
AC
Transmit
Designation
Queue
(informative)
AC_BK
BK
Background
-
AC_BK
BK
Background
0
BE
AC_BE
BE
Best Effort
3
EE
AC_BE
BE
Best Effort
4
CL
AC_VI
A_VI
Video (alternate)
5
VI
AC_VI
VI
Video (primary)
6
VO
AC_VO
VO
Voice (primary)
7
NC
AC_VO
A_VO
Voice (alternate)
Table 4.1 UP-to-AC mappings with 6 transmitting queues
Graceful degradation
The Intra-Access Category Priority element also includes a Drop Eligibility Indicator (DEI)
that is connected to a stream. When a frame with the DEI bit activated is to be transmitted, the
retransmission procedure keeps two different drop-eligible retry counters, long and short, that
may be shorter than the normal ones. Therefore, in the case of bandwidth shortage a stream
that has the DEI bit activated is more likely to reach the maximum number of retransmissions
and be discarded. Also, the DEI bit appears in the MAC frame header and this allows the
frame to be dropped in the receiving station if it has insufficient resources to process it. The
details of when a frame is selected to be dropped are not set in the standard and are left open
to the implementation.
4.1.3
Stream Classification Service (SCS)
Stream Classification Service (SCS), introduced by the TGaa, can be used to classify a range
of traffic classifications to a particular Intra-Access Category Priority element. The SCS
enables the establishment of a classification using layer 2 and/or layer 3 signaling to match
37
Stream Classification Service and Interworking with 802.1AVB
incoming unicast MSDUs. Once classified, unicast MSDUs matching the classification are
assigned to an access category and are tagged with their drop eligibility. When intra-access
category prioritization is enabled, SCS allows MSDUs matching the classification to be
assigned to the primary or alternate transmit queues [9]. Using the combination of the primary
and alternate queue and the Drop Eligibility Bit, a transport stream that is set up between the
station and the AP can achieve better prioritization between other streams.
4.1.4
Related Work
Various attempts have been made in order to provide differentiation mechanisms between
traffic flows within the same priority class, thus reduce the contentions between those flows.
Therefore, much of this work proposes enhancements to DCF and EDCA.
Some of the proposals consist of dynamic adjustment of the CW size. In [23], the authors try
to improve the protocol capacity of IEEE 802.11 networks, by tuning its backoff algorithm.
Their purpose is to show that by observing the network status, it is possible to estimate the
average backoff window size that maximizes the throughput. With this estimation the optimal
value of the CW can be computed, for a given congestion level, without increasing the
collisions in an IEEE 802.11 network. [24] proposes a method based on Kalman filter for
estimating the number of active hosts to set suitable values for CW. The Kalman filter is an
algorithm which operates recursively on streams of noisy input data to produce a statistically
optimal estimate of the underlying system state. The authors also enhance this approach with a
change detection mechanism that detects network state variations and accordingly feed the
Kalman filter. Those solution, however, are hard to be directly used in real wireless
environments due to their high complexity.
In [25], the authors proposed a simple adaptive mechanism, called Dynamic Optimization On
Range (DOOR ) that aims at improving the IEEE 802.11 DCF by integrating system
measurement with adjusting parameters, dynamically. The proposed mechanism offers a low
algorithm complexity and low system cost and it is, also, applicable in complex wireless
networks. This solution, however, cannot be applied to the distributed Ad Hoc mode due to
the need of a central control. Qiang Ni et al. in [26] have introduced a method that is based on
slow CW decrease, that after a successful transmission the CW size is divided by two, instead
of being reset to the minimal value (CWmin). That way the CW value should be kept the
same as long as the congestion level is not likely to drop sharply, providing more collision
38
Stream Classification Service and Interworking with 802.1AVB
avoidance during congestion periods. This method, however, lacks in monitoring channel load
and taking in consideration issues such as the presence of hidden nodes. Thus, it has limited
adaptation to network conditions.
[27] proposed an enhanced service differentiation mechanism called Differentiation Service
based on Per Queue (DSPQ). DSPQ offers an improved ability to differentiate QoS
provisioning between traffic classes, while remaining overall bandwidth efficient. Through
adopting traffic conditions at the entrance of MAC AC queues and maintain collision rate for
each AC, DSPQ dynamically adjusts to changing conditions, confirming oscillations in
throughput and delay. This allows the mechanism to provide strict service differentiation and
good flow fairness, while still maintaining a high level of channel utilization. The simulations
showed that, indeed, DSPQ improves the goodput and reduces delay and collision rate.
Furthermore, since there is no need for a central coordinator, it can also be applicable to QoS
control in wireless Ad Hoc networks.
4.2 Interworking with 802.1AVB
4.2.1
IEEE 802.1AVB
IEEE 802.1 is a working group of the IEEE 802 project of the IEEE Standards Association, an
organization within IEEE that develops standards in a broad range of industries. The IEEE
802.1 working group is concerned with and develop standards and recommended practices in
the following areas: 802 LAN/MAN architecture, internetworking among 802 LANs, MANs
and other wide area networks, 802 link security, 802 overall network management, and
protocol layers above the MAC and Logical Link Control (LLC) layers [28].
The 802.1 working group has four active task groups: Interworking, Security, Audio/Video
Bridging (AVB), and Data Center Bridging (DCB) [28]. In the following pages we will make
a presentation of the AVB task group and its interworking with the IEEE 802.11aa task group.
The IEEE 802.1 AVB task group is developing a set of standards to allow the transport of
high-quality low-latency streaming of time-sensitive audio/video applications over IEEE 802
bridged LANs. Its specific goals include [28]:

Layer 2 time synchronizing service that is appropriate for the most stringent
requirements of consumer electronics applications.

Definition of an admission control system that allows bridges to guarantee the
resources needed for AV streams.
39
Stream Classification Service and Interworking with 802.1AVB

Enhancement of standard 802.1 bridge frame forwarding rules to support AV streams.

Establishment of a set of usage-specific profiles using the AV bridging specifications.

Creation of a standard for AV bridging between specific 802 LAN configurations and
“802-like” LANs.
These enhancements are implemented using relatively small extensions to standard Layer 2
MACs and bridges. This allows non-AVB and AVB devices to communicate with each other
using standard 802 frames. However, as shown in Figure 4.4, only AVB devices are able to:

Reserve a portion of network resources through the use of admission control and
traffic shaping.

Send and receive the new timing-based frames.
Currently, AVB consists of the following four active projects:

IEEE 802.1AS: Timing and Synchronization for Time-Sensitive Applications

IEEE 802.1Qat: Stream Reservation Protocol

IEEE 802.1Qav: Forwarding and Queuing for Time-Sensitive Streams

IEEE 802.1BA: Audio Video Bridging Systems
IEEE 802.1Qat and 802.1Qav are amendments to the base IEEE 802.1Q document, which
specifies the operation of “Virtual Bridged Local Area Networks”. Note that a bridged
network refers to multiple network segments connected to each other at the data link layer
(Layer 2) of the OSI model.
Figure 4.3 Overview of IEEE 802.1 AVB
40
Stream Classification Service and Interworking with 802.1AVB
Figure 4.4 AVB Connections
4.2.1.1 IEEE 802.1AS
IEEE 802.1AS specifies the transport of timing and precise synchronization in bridged fullduplex IEEE 802.3 and IEEE 802.11 networks. This precise synchronization has two
purposes:
i. To allow synchronization of multiple streams.
ii. To provide a common time base for sampling/receiving data streams at a source device
and presenting those streams at the destination device with the same relative timing.
A bridge or end station that meets the requirements of IEEE 802.1AS is referred to as a “timeaware” bridge or end station, respectively. IEEE 802.1AS relies on the transfer of timestamps
using mechanisms that are medium depended. Specifically, for a full-duplex Ethernet
medium, 802.1AS uses a subset of IEEE 1588, a protocol used to define precise timing,
referred to as Precision Time Protocol (PTP). For IEEE 802.11 links, 802.1AS uses timing
facilities, developed initially for location determination (defined in IEEE 802.11v) [29].
An 802.1AS network timing domain is formed when all devices follow the requirements of
the 802.1AS standard and communicate with each other using the IEEE 802.1AS protocol.
Within this domain there is a single device, called the “Grand Master”, that provides a master
timing signal as demonstrated in Figure 4.5. This device can be either auto-selected or can be
specifically assigned. Typically, AVB devices exchange capability information on physical
link establishment, so that network synchronization capable devices will start to exchange
clock synchronization frames. If there are AVB devices with no such capability, an AVB
timing domain boundary is determined (Figure 4.4).
41
Stream Classification Service and Interworking with 802.1AVB
The requirements of IEEE 802.1AS were chosen to provide for low-cost bridges, while still
allowing application performance requirements to be met. IEEE 802.1AS has very few userconfigurable options, consistent with the goal of AVB networks being plug-and-play [29].
Figure 4.5 Clocking hierarchy
4.2.1.2 IEEE 802.1Q
IEEE 802.1Q is the networking standard that supports Virtual LANs (VLANs) on an Ethernet
work. A VLAN is the mechanism that allows the creation of groups of logically networked
devices that act as if they are on their own independent network, even if they share a common
infrastructure with other VLANs [30]. The IEEE 802.1Q standard defines a system of VLAN
tagging for Ethernet frames and the accompanying procedures to be used by bridges and
switches in handling such frames (Figure 4.6). The standard continues to be actively revised
and, as mentioned earlier, two of its revisions (802.1Qat and 802.1Qav) are part of the AVB
set of technical standards.
Figure 4.6 Insertion of 802.1Q Tag in an Ethernet frame
42
Stream Classification Service and Interworking with 802.1AVB
4.2.1.2.1 IEEE 802.1Qat
IEEE 802.1Qat, commonly known as Stream Reservation Protocol (SRP), deals with
reservation and maintenance of QoS for a multimedia flow. It deals with up to seven hops
between the source and the sink, with up to two of the hops being IEEE 802.11 [31]. IEEE
802.1Qat allows network resources to be reserved for specific traffic that is traversing a LAN.
It also provides the necessary QoS by allocating and maintaining buffers within the switches
and/or bridges [32]. Those network resources regard both the end nodes of the data stream and
the transit nodes along the path. Once the reservation is set up, periodic QoS maintenance
reports are monitored to ensure that the required QoS is maintained [31]. An end-to-end
signaling mechanism to detect the success or failure of the effort is also provided.
Many vendors and users desire a single network infrastructure to carry various multimedia
applications such as digital video, high-fidelity digital audio, and gaming traffic, as well as
non-time-sensitive traffic (e.g., data traffic). The application of current IEEE 802 technologies
for high quality time sensitive streaming allows users to load their networks unknowingly to
the extent that the user experience is negatively impacted. To provide the robust guaranteed
QoS capability for streaming applications, the availability of network resources along the
entire data path must be assured before transmission takes place [28]. This requires a signaling
mechanism, called Multiple Registration Protocol (MRP), to be used between the bridge
nodes to complete the resources reservation process.
4.2.1.2.2 IEEE 802.1Qav
IEEE 802.1Qav utilizes methods described in IEEE 802.1Q to separate time-critical and non
time-critical traffic into different traffic classes. It mainly focuses on three building blocks:
traffic mapping or remapping into classes, shaping, and queue management according to class
[33].
Egress port buffers are separated into different queues, each allocated to a specific class. This
ensures a separation of low priority traffic from high priority traffic [34]. AV bridges reserve
priorities 4 and 5 for video and audio traffic, respectively. To guarantee that the highest
priorities are assigned only to the AV streams, it is required that the priority 4 and 5 legacy
traffic that enters the AVB cloud have its priorities remapped. An alternative to this solution is
to increase the number of queues by two, to provide distinct traffic classes for AV uses of
43
Stream Classification Service and Interworking with 802.1AVB
priorities 4 and 5. This approach does not require any change to the number of priorities, as
streams are recognizable by the existence of a reservation [33].
Moreover, all egress ports have a credit-based shaping mechanism to prevent bursty behavior
[32]. At the extreme, the burst size can be reduced to the nominal frame size, and in this case
the frame size becomes fixed and the interval between consecutive frames can be strictly
enforced [33]. As for the queue management, IEEE 802.1Qav uses the timing derived from
IEEE 802.1AS. Finally, dynamic release of reserved critical traffic class bandwidth is also
supported, so lower priority traffic classes without maximum bandwidth allocation can fill
those gaps [34].
4.2.1.3 IEEE 802.1BA
The purpose of this standard is to specify defaults and profiles that manufacturers of LAN
equipment can use to develop AVB-compatible LAN components, and to enable a person not
skilled in networking to build a network, using those components, that does not require
configuration to provide transporting time-sensitive audio and/or video data streams.
Specifically, this standard satisfies needs such as:

The selection of default operating parameters, when the performance requirements of
AVB over various media prevent the use of some portions of other standards. These
parameters must be defined in order to meet the needs of the users of components built
to those standards.

The detection of non-AVB equipment, so that the performance of AVB equipment can
be maintained.

The definition of the configuration parameters of various 802.1 standards, in order to
achieve automatic configuration of AVB networks.
4.2.2
IEEE 802.11aa: Interworking with 802.1AVB
The IEEE 802.11aa Task Group worked closely with the AVB Task Group in order to make
IEEE 802.11 networks compatible with SRP. Changing the standard to provide for this,
required only minor modifications to the MAC layer. The main one was to add a mechanism
that allows for an AP to initiate the Transport Stream setup procedure, whereas it was only
possible for a Transport Stream to be setup when a station requested it. This was necessary for
an IEEE 802.11 network to be able to complete the process of setting up a path between a
talker and a listener as described in the IEEE 802.1Qat standard. A Higher Layer Stream ID is
44
Stream Classification Service and Interworking with 802.1AVB
included in all the messages exchanged when setting up a stream using the AP initiated setup,
which is used to identify the SRP stream.
Also to facilitate the interworking of IEEE 802.11 with SRP networks, it was important to
standardize the mapping between the User Priorities in 802.1D with the access categories and
transmit queues used by the IEEE 802.11 QoS services, which was done together with the
modifications for the SCS service, whether the new alternate queues are used or not.
45
Chapter 5
Multicast Issues and IEEE 802.11aa
In today's wireless networks most applications make use of unicast mechanisms to send data
and voice traffic directly from one point to another. However, there are many instances where
multicasting offers many advantages, such as in streaming a video to multiple users in a
public hotspot. Multicast transmission allows for the conservation of the bandwidth by
reducing unnecessary packet duplication as well as, making use of the inherently broadcast
nature of the wireless medium.
However, there are many issues that must be addressed, regarding multicast communication in
IEEE 802.11 wireless LANs. A basic one is that of guaranteeing reliability. Though the IEEE
802.11 standard does not include this, various efforts have been made to solve this issue.
Some of those proposals are being presented here, classified according to the TCP/IP protocol
suite. Although there are some proposals based on the PHY, such as dual tones based on
additional RF channels, their implementation is very difficult and inefficient. Thus, in the
following sub-chapters only solutions based on layers above the PHY will be presented.
5.1 MAC Layer Solutions
As the wireless medium is inherently unreliable, in the IEEE 802.11 MAC protocol the
reliable transmission of data is guaranteed by the feedback mechanism where every
transmitted frame is acknowledged by the receiver. This mechanism is not used in
multicasting, because receiving acknowledgments from all the receivers would be inefficient,
as it incurs a large overhead and raises issues with the scheduling and synchronization of
receiving them. Additionally the MAC layer of IEEE 802.11 networks specifies that multicast
frames can only be transmitted using the basic access procedure. Therefore, it cannot use the
RTS/CTS mechanism to protect the broadcast of the frame, and this leads to more collisions.
This means that multicast transmission is unreliable, as the source does not know if the data
was not received because of a collision or because of other problems that are common in
wireless communications. Also there is no provision in the MAC layer for retransmissions of
the data to ensure the reliable receipt. It must be implemented by higher layers, which
introduces a large overhead. It would be much more efficient to support reliable multicast
transmission in the MAC layer.
46
Multicast Issues and IEEE 802.11aa
Another problem is that the IEEE 802.11 MAC specifies that multicast frames must be
transmitted using one of the bit-rates in the Basic Rate Set (BRS), which is a minimum set of
bit-rates that must be supported by all stations in a IEEE 802.11 wireless LAN to ensure they
can receive control frames. Although it is not necessary, most Access Points today use only
the lowest bit-rate for transmitting multicast frames. This leads to lower throughput, which is
also worsened by the performance anomaly problem of IEEE 802.11 wireless LANs. This is
the phenomenon where one station transmitting in low bit-rate causes a large degradation to
the throughput of the whole network, as the low bit-rate transmission occupies the medium for
a significant amount of time. For all these reasons, transmitting multicast frames in legacy
IEEE 802.11 cannot support a reliable, high rate multicast stream. The problems with
multicast in IEEE 802.11 networks have been studied in many previous works. Most of the
proposed solutions can be divided in two categories.
5.1.1
Negative Feedback-Based Solutions
The first category is based on negative feedback (NFB-based). Kuri and Kasera [35] presented
the Leader-Based Protocol (LBP) for reliable multicast over wireless LANs. According to this
protocol, one chosen leader is responsible for sending CTS and ACK frames responding to
RTS and data packets, respectively. On erroneous reception of a packet, the leader does not
send an ACK, prompting the sender to retransmit the packet. On erroneous reception of a
packet at receivers other than the leader, the LBP allows negative acknowledgments (NACKs)
from these receivers, which will collide with the leader's ACK and result in destroying the
acknowledgment and prompt the retransmission of the packet from the sender.
However, leader-based protocols face several problems that exist in all NFB-based protocols,
such as type-unknown for lost packet and frame aggregation support [36]. The first one could
happen when collisions or link errors occur. At these situations the group member does not
receive the frame correctly, thus cannot acquire information contained in the MAC header,
regarding the source and destination address. As a result, it is difficult for the receiver to
decide which node the feedback should be sent to. As mentioned, a collision between an ACK
and a NACK will lead to a retransmission. However, when frame aggregation is considered,
the sender will retransmit the whole aggregated unit even though not all frames in that unit
were transmitted unsuccessfully. LBPs have to deal with other major problems as well, such
as unnecessary retransmission that happens when some receivers send NACK for an
erroneous retransmitted frame that they received, regardless of whether this frame has been
47
Multicast Issues and IEEE 802.11aa
received successfully before or not. This result in redundant retransmission. Generally, LBPs
could not be considered perfectly reliable as the only leader cannot provide feedback about
the rest of the receivers.
5.1.2
Positive Feedback-Based Solutions
The second category is based on positive feedback (PFB-based). Kuri and Kasera [35], also
presented the Delayed Feedback-Based Protocol (DBP) which differentiates from LBP mainly
in two ways. The first is that MCTS frames are sent by each receiver instead of only the
leader. The second difference is that in order to avoid the collisions between NACKs and
MCTSs, the receivers will have to wait for a random time before transmit a MCTS. Another
issue of DBP is the choice of the right parameter for waiting times.
Tang and Gerla proposed the Broadcast Media Window (BMW) [37] protocol that supports
reliable multicast in ad hoc networks. The basic idea in this scheme is to treat each multicast
request as multiple unicast requests by transmitting a sequence of frames in each round of
transmission to each receiver in turn. When channel contention is high, this protocol becomes
inefficient and reverts back to the unreliable delivery of IEEE 802.11. The Batch Mode
Multicast MAC (BMMM) proposed by M.T.Sun et al. [38] is based on the BMW protocol but
reduces the number of contention phases when the acknowledgments are sent to make it more
efficient. At this approach the sender sends RTS frames to each station of the multicast group
and waits for CTS replies from each of them (Figure 5.1). Upon successful reception of the
CTSs the sender transmits the data frame and then it sends a Request for ACK (RAK) frame
to each station individually (Figure 5.2). All the receivers who successfully received the data,
reply with an ACK. Upon receipt of ACKs from all the recipients, the transmission is
completed. If there were stations that did not reply with ACK, the sender transmits the data
frame again, except this time only to subset of stations whose ACKs where not received.
Although BMMM is a rather efficient approach to achieve reliable multicast, it still has a very
high control traffic overhead (Figure 5.3).
Figure 5.1 Primary idea of BMMM
48
Multicast Issues and IEEE 802.11aa
Figure 5.2 RAK frame format
Figure 5.3 BMW vs BMMM
In [36], the Double Piggyback Mode Multicast (DPMM) protocol is presented trying to
decrease the control packet overhead and improve the efficiency by embedding the
acknowledgment information of a received data frame, in the next CTS frame. Lyakhov et al.
[39] proposed the Enhanced Leader Based Protocol (ELBP) that further improves the BMMM
protocol by using the BlockAck mechanism from IEEE 802.11e. Another improvement of the
BMMM is proposed in [40]. V. Srinivas and L. Ruan presented the Slot Reservation Based
(SRB) reliable multicast protocol. The SRB uses the RTS-CTS-DATA-ACK exchange
mechanism to ensure reliability. Upon successful association with a station, the AP sends,
among others, two parameters called Association ID (AID) and Multicast Association ID
(MAID) (Figure 5.4). This helps to establish a schedule of transmissions of the multicast
receivers and avoid the collisions that otherwise would be occurred by the simultaneous
transmission of the stations' ACKs and CTSs. To ensure efficiency in retransmissions, the
SRB uses a modification of the RTS multicast frame that is sent out at the beginning of each
retransmission. The RTS frame is appended with bitmap with n bits, where n is the number of
the stations in multicast group and each bit corresponds to a station, resulting that way in a
one-to-one mapping. In any retransmission phase only the bits that correspond to the
participant stations are set to 1. With this mechanism the data are being retransmitted only to
49
Multicast Issues and IEEE 802.11aa
those stations from whom the AP did not receive ACK, improving significantly the efficiency
and throughput (Figure 5.5) .
Figure 5.4 Comparison of the initial transmission phase
Figure 5.5 Comparison of the retransmission phase
5.2 Network Layer Solutions
Though, most of the proposals for reliable multicast transmission are based on the MAC layer,
there are a few others that are applied on the upper layers. For multicast on the network layer,
two interesting approaches have been made, the first of which gives a solution for P2P live
streaming multicast transmission and the second for video multicast transmission.
In P2P live streaming systems the data are being exchanged between peers by application
layer multicast protocols, which in turn usually use unicast transmission in the network layer.
This results in a lot of bandwidth waste. To overcome this problem, Wenbin Jiang et al. [42]
proposed a wireless multicast agent mechanism (WiMA) that uses IP multicast, reducing that
way the network overhead and, thus, being able to support more users. The main idea behind
this is to select a multicast agent that acts as a proxy to the other wireless peers. It is also
necessary to set one multicast IP address for each channel. The selection of the agent is a
process that takes in consideration parameters such as radio signal, network traffic and power
50
Multicast Issues and IEEE 802.11aa
of the CPU of the candidate agent. Though this solution does reduce the bandwidth
consumption, it does not include an efficient QoS mechanism.
Another attempt for reliable video multicast transmission on IEEE 802.11 WLAN has been
done in [43]. There, Maria Samokhina et al. proposed a scheme based on raptor code that is
applied right above the MAC layer. According to their proposal, the AP before the
transmission of data first determines the expected symbol loss rate by taking in account the
channel status, and then using the raptor encoder, encodes the data and generates as many
encoded parity symbols, as need to prevent the losses. The IEEE 802.11n Power Save MultiPoll (PSMP) burst of two rounds is used, for the transmission. During the first round, the AP
collects information based on the Singal-to-Noise Ratio (SNR) feedback from the multicast
receivers to calculate the raptor symbol loss rate. In case that the SNR is different between
the receivers, the AP takes in account the worst SNR value. The receivers send their feedback
to the AP by sending Reliable Multicast (RM) and Feedback (FB) frames. These frames are
shown in Figure 5.6. Even if one FB frame is not received by the AP, the process starts again.
This does not increases significantly the network overhead because the probability that an RM
or FB frame fails is much smaller than that of a data frame failure due to their small size. The
multicast data transmission takes place during the second round of PSMP burst, where each
raptor symbol forms an MSDU (Figure 5.7). The main disadvantage if this proposal is that it
only supports video multicast transmission.
Figure 5.6 RM and FB frame formats
51
Multicast Issues and IEEE 802.11aa
Figure 5.7 An 802.11n A-MPDU based scheme
5.3 Transport Layer Solutions
In [44], H. Fujisawa et al. examined the multicast packet loss due to collisions with TCP
packets on IEEE 802.11 WLANs and presented two significant results. First, they proved
after simulation that the packet loss rate does not increase as the number of TCP station
increases. Actually, they stated that the ratio remains constant if the number of the stations is
more than 3. That happens because TCP's flow control, which uses ACKs, prevents the
stations from contending. Moreover, they stated that as the multicast transmission rate
increases, the packet loss rate decreases. The second result is that the multicast packet losses
can be recovered by a FEC technique, the Reed-Solomon (RS) code which shortens the code
length to improve error correcting capability.
5.4 Application Layer Solutions
In [45], M. Kappes et al. presented the Application Layer Communicator (ALC), an
application-layer approach for ad-hoc networks. The ALC uses MAC/IP multicast without the
need to apply an changes to the MAC or IP layer. It also uses UDP as its transport protocol, at
which some further headers need to be added as illustrated in Figure 5.8. Where the
immediate source address is the MAC address of the last device that forwarded the frame.
Each ALC node periodically sends out a list of its neighbors just like in a distance vector
routing protocol. By keeping a list of all stations, each node can determine the next hop in an
efficient route. This approach does not declare a mechanism that calculates the retransmission
timers.
52
Multicast Issues and IEEE 802.11aa
Figure 5.8 ALC frame format
R. Chandra et al. proposed in [46] proposed an application layer multicast technique that takes
advantage of the multi-access nature of the wireless medium. This mechanism requires a
DirCast server, as shown in the figure Figure 5.9. Also, each multicast receiver must have
installed a DirCast client software. The idea behind this is that the server sends multicast
packets as unicast to a selected station at each AP. The main goal now is to select the
appropriate target client. As the AP adjusts the transmission rate based on the client's channel
conditions, it is logical that the worst client has to be selected so that all the receivers can
decode successfully the received packets. The DirCast re-calculates the data rate that the AP
uses every 30 seconds. Moreover, this value is re-evaluated every time a new client joints and
whenever the loss rate at any client exceeds the 10%.
Figure 5.9 DirCast operational overview
C. Bravo and A. Gonzalez [47], proposed a polling-based service to improve the reliability of
IEEE 802.11 wireless LANs for multicast data. Here, a Smart Server is required for each AP
53
Multicast Issues and IEEE 802.11aa
in a network. A Smart Client mechanism, placed right below the client application, is also
required (Figure 5.10). The process starts with the server transmitting the first 8 frames. If one
of those is lost, all the others with sequence number greater than this of the lost one, are stored
in a buffer until the frame is retransmitted and received successfully or until the buffer is filled
up at a certain point where the lost frame is ignored and the all of the waiting frames are sent
up to the client application. Once the server has completed the transmission of the frame
group, it sends out a poll message asking for the reception status of the frames of that group.
The Smart Client, then, set the Bit Array of the poll message depending if it successfully
received or not the frames of the group. At the next round the Smart Server after having
received the Bit Array, transmits the next group of frames, including those that were not
received by the stations, previously.
Figure 5.10 GP protocol network diagram
54
Multicast Issues and IEEE 802.11aa
Figure 5.11 GP protocol over time diagram
In [48], a Cooperative Peer-to-peer Repair (CPR) strategy that allows peers in Wireless Wide
Area Networks (WWANs) to cooperatively repair lost packets from neighbors watching
different views, is presented. To succeed this, the server should transmit dept maps in addition
to texture maps, so that a viewer A watching a video can repair a lost frame of viewer B, by
synthesizing in B's viewpoint its frames via depth-image based rendering (DIBR). To do so,
the two viewers need to be watching the video in different views.
5.5 Cross-Layer Solutions
There have, also, been proposed other solutions that are not based on a single layer, but
instead they need cross-layer information to be exchanged. Those solutions can be categorized
based on the network topology as centralized, where an AP manages the network traffic, and
decentralized, where the network nodes do not rely on any preexisting infrastructure.
5.5.1
Decentralized Networks (Ad-Hoc Networks)
In [49], there has been implemented a scheme that relies on the cooperation of both the MAC
and application layer and aims to improve the video quality, thus to minimize the distortion, at
all network nodes. To do so, each node consists of a link state monitor at the MAC layer and a
55
Multicast Issues and IEEE 802.11aa
video rate controller at the application layer (Figure 5.12). By sending periodic broadcasts, all
the nodes are able to collect information regarding their neighbors' link utilization and
congestion values. Based on that, an accumulated price is calculated that is being forwarded
up to the application layer, where, in turn, the controller uses it to calculate the appropriate
video rate. Moreover, the receiving rates may vary between nodes that are accepting the same
video, because of the heterogeneity that might be in their link speeds. Thus, rate adaptation
needs to be implemented, and this is being accomplished by adopting the Scalable Video
Coding (SVC) extensions of the H.264/AVC standard. Although, the experimental results
have shown that it is pretty effective, this solution has the strong limitation that it is not
applicable on transmissions that carry various data types.
Figure 5.12 Each node consists of two layers
Another interesting attempt to improve multicast transmissions in ad hoc networks has been
presented in [50]. Soon Y. Oh et al. proposed the MIMO-Cast protocol, that works in
conjunction with Multi Point Relay (MPR) and, basically, consists of two mechanisms. The
first is that it builds a multicast tree of all nodes, similarly to the ODMRP protocol, with the
difference that only the selected MPR nodes can forward the message. That way, the network
overhead is reduced significantly (Figure 5.13). The second mechanism takes care of
eliminating completely the problem of the hidden node; as each terminal transmits at different
radio weight, selective reception is applied and thus, interference signals are being blocked.
Simulation results have shown that MIMO-Cast performs far better than conventional
multicast protocols with IEEE 802.11.
56
Multicast Issues and IEEE 802.11aa
Figure 5.13. Normal flooding (A) vs proposed scheme's flooding (B)
C.C. Hu [51] presented another cross-layer scheme that works with the cooperation of the
physical, the MAC and the network layer, as well and aims at bandwidth consumption
improvement and data rate selection for multicast transmissions in multi-rate Mobile Ad-Hoc
NETworks (MANETs). In the proposed protocol, the hosts monitor the channel's status, that
can be either idle or busy. Also, when a node is forwarding data frames, its neighbors are
blocked and cannot transmit anything within this period. The goal here, is to calculate the
appropriate data rates, based on the number of each forwarder's neighbors, to improve the
network performance. That means that there must be selected a tree that consists of forwarders
with as fewer neighbors as possible. The tree determination starts from the server by
broadcasting a network layer control packet. When a node receives that packet, it responds
with another control packet that carries information about the link time with its neighbors, the
rest of its bandwidth and the relations with the neighbors that is kept in its MAC layer. The
server, then, can determine the best tree and data rates, using this information. If a client
detects a change that renders the tree to be considered outdated, then it broadcasts another
control packet, so that, the server can update the tree.
5.5.2
Centralized Networks (Access Point-Based Networks)
Jose Villalon et al. [52] introduced the Auto Rate Selection mechanism for Multicast
(ARSM), that dynamically adapts to the channel conditions changes and selects the
appropriate data rate. This process is done with the collaboration of the PHY and MAC layer.
The main idea is that for each multicast group is being selected a leader. First, the AP sends
out to the group a Multicast Probe (MP) frame. When the members receive that frame, they
calculate the Signal-to-Noise Ratio (SNR) value of the channel. Based on this value, each
station estimates the point in time that it will respond to the AP. The station with the lowest
SNR value, that is the worst channel, responds first by sending a Multicast Response (MR)
57
Multicast Issues and IEEE 802.11aa
frame to the AP. MP and MR frames are represented in Figure 5.14. This node will then be
selected as the group leader and will be responsible for acknowledging the received multicast
data on behalf all the other members, which, however, may issue NACK frames in case they
detect errors in the received frames. The AP in this case retransmits the data.
Figure 5.14. Control frames formats
Although, this mechanism guarantees reliability, it has the disadvantage that the data rate
adaptation is made based on the station with the worst channel conditions, penalizing that
way, the rest of the stations that are able to communicate at higher rates. The need of a
mechanism that also adapts to this heterogeneity, led the same team to a cross-layer
architecture that extends their first approach of ARSM to the application layer, as well [53].
This enhanced version of ARSM, called Hierarchical ARSM (H-ARSM), provides further
services that allows the nodes with better channel conditions to take advantage of that, at least
at a certain point. The H-ARSM is only implemented for video communications. With this
mechanism implemented, the video is decoded into two layers; the base layer, which is
supposed to provide the minimum acceptable video quality, and the enhancement layer. The
frames that contain the data for the base layer are transmitted to all the members of the
multicast group, according to the rules defined by the ARSM. The frames that contain the data
for the enhancement layer follow an operation similar to that of the ARSM with a main
difference, though. Instead of the leader being the station with the worst channel, now this
selection is being made based on the best SNR. So, in this case the station with the best
channel conditions has the responsibility of acknowledging the frames with the enhancement
layer data, on behalf all the members of the group. Again, the other stations may reply with a
NACK upon erroneous reception of the data. The formats of the MP and MR frames are the
same as those used by ARSM. The simulation results have shown that in real-time video
streaming the quality of the video can be improved significantly. However, in a scenario
58
Multicast Issues and IEEE 802.11aa
where exists only one station in the multicast group that can receive data at the high ratio of
the enhancement layer, only this station would be able to receive the video in a high quality
and the rest of them would have to compromise with the minimum acceptable video quality.
A very similar mechanism to the above, is presented by W.S. Lim, D.W. Kim and Y.J. Suh in
[54]. They first presented a MAC-layer protocol for multicast transmissions in IEEE 802.11n
WLANs, called Reliable and Efficient Multicast Protocol (REMP) and then they extended it
via cross-layer optimization, to support scalable video streaming. This enhanced version of
REMP is called Scalable REMP (S-REMP). S-REMP is basically a leader-based protocol,
where the leader is the node with the worst channel conditions. Under S-REMP, when the
channel condition is stable the AP exchanges control frames only with the leader, whereas
when it is dynamic, it exchanges control frames with all the network nodes and based on their
feedback, -that is, each node's channel condition- the AP adjusts accordingly the modulation
and coding scheme (MCS). S-REMP's goal is to provide high-quality video to the users with
high SNR value when the available bandwidth is not enough for the AP to transmit the entire
video to all the multicast members. To do so, each A-MPDU is separated into two sub AMPDUs, when the network overhead becomes heavy. The first sub A-MPDU contains the
base layer of the video and is transmitted to every network station with a low MCS, to
guarantee the minimum video quality to all, while the second sub A-MPDU that contains the
enhanced video layer is transmitted with a higher MCS to avoid packet drops.
To better manage the multicast group, the AP maintains a GroupTable of eleven fields. Group
address field and leader address field indicate the multicast group address and leader address,
respectively. The address list field contains the addresses of all the receivers, while SNR list
field contains their SNR values. The timestamp field keeps the time of the last transmission
for the group. Tdelay field stores the last waiting time for channel access. Drop field
represents whether one or more data frames where dropped since the last data transmission.
This field is set to 1 every time a frame destined for the multicast group is dropped due to
overload at the AP's data queue and is reset to 0 for each data transmission. LoadUp and
LoadDown fields indicate the network load status and are updated before each data
transmission, according to the drop field. K field represents the number of the enhancement
layers that are included in the second sub A-MPDU and mAlg2 field is a MCS value for the
second sub A-MPDU.
59
Multicast Issues and IEEE 802.11aa
S-REMP has the advantage over H-ARSM [53] that the enhancement layers are not
transmitted at a ratio that is based only on the SNR value of the station with the best channel
condition. Instead, under S-REMP the MCS selection for such layers, takes in account other
parameters such as the K field and the MCS value of the first sub A-MPDU.
O. Alay et al. [55] proposed an alternative solution that is implemented both in the MAC and
application layer. The main idea here, is to dynamically adapt the data transmission rate and
the Forward Error Correction (FEC), as well, in order to succeed the maximum video quality
for all the multicast members. In multicast transmissions, expecting the AP to retransmit the
lost frames to all the multicast members, obviously is not a good idea as it may lead to
extremely high network overhead. FEC implemented in the application layer, on the other
hand, is a good alternative solution for handling erroneously received packets. For this
solution to be more effective, CRC-based error detection needs to be implemented at the link
layer, so that any corrupted packets would first be removed and then recovered by the
application layer's FEC. This solution however adds some overhead at the network since more
parity packets are now needed. Moreover, the count of those packets increases as the Packet
Error Rate (PER) increases, too. Therefore, it is important to always know the PER of the
multicast group, in order not to apply more FEC parity packets than needed. The answer to
this issue, is that all the multicast receivers should periodically send PER information to the
AP. Using this feedback, the AP adapts the transmission rate and the FEC based on the worst
station's PER. From experiments they ran, the authors concluded that using higher
transmission rate, thus more FEC information, is more efficient than using lower transmission
rate along with weaker FEC. Based on this, they focused on using the higher transmission rate
possible together with the appropriate FEC. The rate adaptation method that is adopted in the
current thesis, is that the optimal transmission rate is chosen by searching through different
rates. The resulting algorithm has two major features.
The first is to switch the transmission rate. When the PER is 0% - 15%, they switch to a
higher transmission rate. This means that if, for example, for a rate of 1 Mbit/s the PER was
less than 15%, then the next transmission will be at 5.5 Mbit/s. This also means, that the next
PER will be less than 40%, which is the worst case scenario. If at any time the PER is more
than 40%, then it is considered that the current transmission rate cannot be sustained. In this
case, the mechanism switches at the base rate, which is 1 Mbit/s. PER higher than 50% was
not studied, since in such a case there was observed frequent network connection lost. The
60
Multicast Issues and IEEE 802.11aa
second component of the proposed scheme is to keep the transmission rate stable. This
happens in case the PER is more than 15% while transmitting at the base rate. Jumping to the
next rate, in this case is not worth as the PER may become higher than the 40%.
While the transmission rate and FEC switch, the video rate switches, as well. It also switches
even if the transmission rate remains stable because the FEC is always being adjusted based
on the received PER. Basically, the video rate is adapted in the following way; regarding the
transmission and FEC rate, the video is pre-encoded in different bit rates, resulting in various
versions of the same video. To simplify the system, the authors suggested to create only two
different video streams for each transmission rate.
Transmission rate
1 Mbit/s
5.5 Mbit/s
11 Mbit/s
PER range (%)
Video rate (Mbit/s)
0-25
0.13
25-40
0.10
0-25
0.70
25-40
0.52
0-20
1.44
20-40
0.98
Table 5.1 Video rate adaptation based on PER for different transmission rates
The simulation results confirmed that this mechanism improves the network performance
when it is about multicast transmission. However, it lacks of an efficient-enough video rate
adaptation mechanism and a more accurate FEC prediction method, as currently only the PER
of the last transmission is used, instead of a history of the reported PERs. Finally, the
proposed solution has not yet been adapted in other environments than the IEEE 802.11b.
5.6 The IEEE 802.11aa Task Group Approach
5.6.1
Group Addressed Transmission Service (GATS)
In order to provide reliable multicast, the IEEE 802.11aa amendment specifies the Group
Addressed Transmission Service (GATS) that allows a station to request greater reliability for
a group addressed stream. The service, in addition to the legacy No-Ack/No-Retry multicast,
comprises the Directed Multicast Service (DMS) and the Groupcast with Retries (GCR).
When setting up a stream to a multicast group with the AP, a station can request to use any of
these policies. The policy that is used for a particular stream can be changed dynamically later
[56].
61
Multicast Issues and IEEE 802.11aa
5.6.1.1 DMS Procedures
This method was first introduced in the IEEE 802.11v amendment for Wireless Network
Management. In IEEE 802.11aa, however this method can be used dynamically and switched
to the GCR policy. The implementation of DMS is optional for a WiFi Multimedia (WMM)
station and mandatory for a robust Audio/Video (AV) streaming station. The DMS converts
multicast traffic to unicast frames directed to each of the group recipients in a series. The
transmission uses the normal acknowledgment policy and will be retransmitted until it is
received correctly. This is obviously the most reliable scheme but it also has the greatest
overhead and does not scale well to multicast groups with a large number of members.
5.6.1.2 GCR Procedures
GCR is a flexible service that aims at increasing the reliability of group addressed frames,
while providing a better scalability of the DMS. GCR may be provided either by an AP to the
associated stations of a BSS, or by a mesh station to the rest of the peer stations in a mesh
BSS. The TGaa group specifies that the GCR service uses the same setup, modification and
termination procedures as the DMS. GCR is an extension of DMS. In particular:

A GCR agreement applies to a single group address, whereas a DMS flow is not
restricted to a single group address.

DMS offers multicast-to-unicast conversion only, whereas GCR includes several
retransmission policies and delivery methods.
GCR extends the DMS Request and DMS Response elements with the addition of the GCR
Request and GCR Response sub-elements respectively, for managing the set up, the
modification and the tear down of GCR services between the GCR service provider, that is an
AP or a mesh station, and the associated group addressed stations. The GCR service has two
delivery methods, regarding group addressed frames:

Non-GCR Service Period (non-GCR-SP)

GCR Service Period (GCR-SP)
GCR-SP transmits GCR group addressed frames at intervals that might be smaller than the
beacon interval. Compared to non-GCR-SP, GCR-SP might provide lower delay and jitter.
Further analysis regarding the functionality of those two delivery methods, is beyond the
scope of this thesis.
62
Multicast Issues and IEEE 802.11aa
Moreover, GCR defines two additional retransmission policies for group addressed frames, in
addition to the "No-Ack/No-Retry" and DMS mechanisms:

GCR Unsolicited Retry

GCR Block Ack
The TGaa also specifies the GCR transmission concealment; a mechanism that prevents group
addressed traffic transmitted via either of the above two retransmission policies, from being
passed up through the MAC layer of GCR-incapable stations.
5.6.1.2.1 GCR Unsolicited Retry
This mechanism allows the transmission of the multicast frames to be repeated a number of
times to increase the probability of correct reception at the stations that are listening to the
particular group address. No acknowledgment mechanism is used. The number of
retransmissions is not specified in the standard; it depends on the implementation and is
subject to applicable frame lifetime and retry limits. Note that the AP or a mesh station may
vary the lifetime limit for a group address at any time and may use different lifetime limits for
different GCR group addresses. The standard also suggests a protection mechanism to be used
to prevent other stations from transmitting at the same time. However, the GCR unsolicited
retry uses a backoff algorithm to avoid collisions, regardless of whether or not a protection
mechanism is in use. To detect and prevent frame duplications, the receiving stations keep a
cache of recently received frames. In general, this method may provide moderate delay,
efficiency and reliability but has high scalability, thus it is better suited for groups with large
number of members.
5.6.1.2.2 GCR Block Ack
This method extends the Block Ack mechanism specified in IEEE 802.11e for use in multicast
transmissions to a group. The AP transmits a number of multicast frames and then requests
from one or more of the recipients to acknowledge the receipt of the transmitted frames, by
sending a modified Block ACK Request (BAR) control frame. Frames that have not been
received correctly by one or more of the receivers can then be scheduled for retransmission.
For this mechanism to function, every recipient of a GCR Block Ack agreement shall
maintain a block acknowledgment record that includes:

a bitmap, indexed by sequence number

a 12-bit unsigned integer starting sequence number
63
Multicast Issues and IEEE 802.11aa

WinStartR, representing the lowest sequence number position in the bitmap

a variable WinEndR

the maximum transmission window size, WinSizeR.
Upon a BAR reception, a station responds with a BlockAck frame containing the above
information that the originator will use to determine which frames to retransmit. Although
each station shall transmit a BlockAck frame at a delay of SIFS after the BAR reception, the
standard suggests a protection mechanism to be implemented in order to avoid collisions.
The choice of the stations from which an acknowledgment will be requested is left to be
decided by the implementation, and therefore this method can be used with many leader-based
multicast methods and algorithms for selecting a leader. This method can offer a high degree
of reliability, scalability and performance, with the trade-off depending on the
implementation. However, the current hardware on the market does not allow to ACK
multicast or broadcast frames, thus it is not possible to arrange such a mechanism over
existing MAC, where packets not requiring acknowledgment are immediately discarded after
transmission.
In the beginning stages of the Task Group discussions it was suggested that a scheduled
BlockAck mechanism would be used, where the AP would send one modified BlockAck
Request with a bitmap specifying the order in which the receivers would answer, and the
receivers would send their acknowledgments in the specified order, separated by SIFS time.
However this was canceled due to several worries such as the effect of air propagation time,
the effect of hidden stations or of a station missing the BlockAck Request. Now only the
Polled BlockAck mechanism [57] is specified, where the AP requests an acknowledgment
from each receiver separately. This does not require any changes to the BlockAckRequest
frame, is easier to handle within hardware and does not cause legacy client compatibility
issues.
Table 5.2 summarizes the characteristics of the different policies specified by the Group
Addressed Transmission Service. The legacy No-Ack/No-Retry multicast and the three
alternative policies are illustrated in Figure 5.15.
64
Multicast Issues and IEEE 802.11aa
Multicast Policy
Overhead
Scalability
Complexity
Reliability
Legacy multicast
None
High
Low
Low
DMS
Large
Low
Medium
High
GCR Unsolicited
Retry
Small
High
Low
Implementation
depended
GCR Block Ack
Implementation Implementation
High
Implementation
depended
depended
depended
Table 5.2 Characteristics Of Group Addressed Transmission Service Policies
Figure 5.15 Group Address Transmission Service with different policies
5.6.2
Other Solutions
Another proposal that was made to the IEEE 802.11aa Task Group is by Miroll and Li [58],
who suggested a leader-based protocol that introduces the idea of feedback cancellation where
a Negative ACK (NACK) frame transmitted simultaneously with the leader’s ACK frame
cancels the acknowledgment at the receiving AP and forces a retransmission (Figure 5.16).
This proposal also includes a FEC coding scheme where additional multicast frames are
transmitted, to allow the receiver to decode the multicast stream even if some of the frames
are corrupted. After discussion, the proposal was not included in the amendment as it required
many modifications to the IEEE 802.11 standard.
65
Multicast Issues and IEEE 802.11aa
Figure 5.16 The proposed NACK based LBP
In [59], the Multicast Collision Free (MCF) scheme is proposed. This mechanism is based on
two steps. First the transmitter sets a multicast flag active to avoid collisions that may arise if
other stations attempt to transmit at the same time. Note that this mechanism does not use
ACK, so frames will not be retransmitted. In the second step, the transmitter has the multicast
flag active and begins the multicast transmission after waiting PIFS making, this way, sure
that the medium is idle. Figure 5.17 demonstrates a flow chart of the MCF mechanism, while
in Figure 5.18 a comparison between the MCF and the standard channel access, is presented.
Figure 5.17 Multicast Collision-Free Mechanism
66
Multicast Issues and IEEE 802.11aa
Figure 5.18 Multicast channel access: (a) Standard, (b) MCF
67
Chapter 6
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
6.1 Industrial, Scientific and Medical (ISM) Radio Bands
Industrial, Scientific and Medical (ISM) band is a part of the radio spectrum that can be used
by anybody without a license in most countries. In the United States, the 902-928 MHz, 2.4
GHz and 5.7-5.8 GHz bands were initially used for machines that emitted radio frequencies,
such as RF welders, industrial heaters and microwave ovens, but not for radio
communications [60].
In 1985, the Federal Communications Commission (FCC), a U.S. government agency that
regulates interstate and international communications, opened up the ISM bands for wireless
LANs and mobile communications. In 1997, FCC added additional bands in the 5 GHz range,
known as the Unlicensed National Information Infrastructure (U-NII). Europe's HIPERLAN
wireless LANs use the same 5 GHz bands, which are entitled the "Broadband Radio Access
Network" [60].
6.1.1
2.4 GHz Wireless Band
The 2.4 GHz band is utilized by the IEEE 802.11 b/g/n standards and is divided into 14
channels, with the last one being used only in Japan and allowed only for IEEE 802.11b. The
first 13 channels are spaced 5 GHz apart, with channel 1 being centered on 2.412 GHz and
channel 13 on 2.472 GHz (Figure 6.1). The 14th channel, added by Japan, is centered on
2.484 GHz, 12 GHz above the 13th. Availability of channels is regulated by country,
constrained in part by how each country allocates radio spectrum to various devices. Each of
the channels has signal width of 22 MHz, resulting in only three non-overlapping channels (1,
6, 11) among the first 13. With the latest IEEE 802.11g standard, however, there are four nonoverlapping channels (1, 5, 9, 13) among the first 13, since IEEE 802.11g uses 20 MHz
signals instead of 22 MHz.
68
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
Figure 6.1 Channels of 22 MHz bandwidth in the 2.4 Ghz band
Channel Center Frequency (MHz)
U.S. (FCC)
Europe (ETSI)
Japan
1
2.412
Yes
Yes
Yes
2
2.417
Yes
Yes
Yes
3
2.422
Yes
Yes
Yes
4
2.427
Yes
Yes
Yes
5
2.432
Yes
Yes
Yes
6
2.437
Yes
Yes
Yes
7
2.442
Yes
Yes
Yes
8
2.447
Yes
Yes
Yes
9
2.452
Yes
Yes
Yes
10
2.457
Yes
Yes
Yes
11
2.462
Yes
Yes
Yes
12
2.467
No
Yes
Yes
13
2.472
No
Yes
Yes
14
2.484
No
No
Only IEEE
802.11b
Table 6.1 WLAN channel frequencies in 2.4 GHz and their availability per region
6.1.2
5 GHz Wireless Band
The 5 GHz band is utilized by the IEEE 802.11 a/n standards and is composed of four
frequency bands: 5.150 – 5.250 MHz, 5.250 – 5.350 MHz, 5.470 – 5.725 MHz and 5.725 –
5.850 MHz. As with 2.4 GHz band, not all 5 GHz band channels are available to every region.
To be more precise: the 5 GHz band has 24 channels in the U.S. and 19 in Europe with 20
MHz bandwidth each, and a further 11 channels in the U.S. and 9 in Europe with 40 MHz
bandwidth each. Operating at the 5 GHz band instead of 2 GHz offers several advantages,
such as better penetration, better scatter and larger number of non-overlapping channels that
results in less radio congestion.
69
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
Channel
Center Frequency (MHz)
U.S. (FCC)
Europe (ETSI)
Japan
34
5.170
No
No
Yes
36
5.180
Yes
Yes
Yes
38
5.190
No
No
Yes
40
5.200
Yes
Yes
Yes
42
5.210
No
No
Yes
44
5.220
Yes
Yes
Yes
46
5.230
No
No
Yes
48
5.240
Yes
Yes
Yes
52
5.260
Yes
Yes
Yes
56
5.280
Yes
Yes
Yes
60
5.300
Yes
Yes
Yes
64
5.320
Yes
Yes
Yes
100
5.500
Yes
Yes
Yes
104
5.520
Yes
Yes
Yes
108
5.540
Yes
Yes
Yes
112
5.560
Yes
Yes
Yes
116
5.580
Yes
Yes
Yes
120
5.600
No
Yes
Yes
124
5.620
No
Yes
Yes
128
5.640
No
Yes
Yes
132
5.660
No
Yes
Yes
136
5.680
Yes
Yes
Yes
140
5.700
Yes
Yes
Yes
149
5.745
Yes
No
No
153
5.765
Yes
No
No
157
5.785
Yes
No
No
161
5.805
Yes
No
No
165
5.825
Yes
No
No
Table 6.2 WLAN channel frequencies in 5 GHz and their availability per region
IEEE 802.11 Standard
2.4 GHz Band
5 GHz Band
802.11a
No
Yes
802.11b
Yes
No
802.11g
No
Yes
802.11n
Yes
Yes
Table 6.3 Radio bands that common IEEE protocols operate in
70
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
6.2 The Overlapping BSS Problem
The overlapping BSS problem refers to situations that two or more systems, unrelated to each
other, are in close enough proximity to hear each other physically [61]. In other words, the
stations or the AP of one BSS are able to receive frames from the other BSS. This is
commonly known as the OBSS problem and is generally considered undesirable, because
members of the OBSSs that interfere with each other and compete for channel access, may
cause increased channel contention level and decreased performance [62]. In particular:

Due to the doubling of the number of stations, the medium contention level increases
dramatically [63].

Interference makes it difficult for a wireless network to provide robust performance
and lead to transient failures [64]. Hence, the stations can not receive the frames
correctly [65].

The expansion of the number of the hidden stations in the overall network system
increases the probability of collisions.
Bellow are some of the possible overlapping scenarios.
Figure 6.2 Scenario 1: Two OBSSs, APs within range of each other
Scenario 1 denotes an overlapping scenario where APs and some stations both from BSS1 and
BSS2 interfere. Additionally, there are some stations from both BSSs, so-called hidden nodes,
that increase the collision probability.
71
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
Figure 6.3 Scenario 2: Two OBSSs, APs not within range of each other
Scenario 2 illustrates two BSSs where the APs are not within the range of each other (Hidden
APs), thus, the number of collisions may increase.
Figure 6.4 Scenario 3: Three OBSSs, APs within range of each other
In scenario 3, an overlapping scenario of three OBSSs is presented, where the AP and some
stations of each BSS are within the range of each other, thus they do listen physically to each
other. Moreover, there are hidden nodes from each BSS resulting in high collision probability.
72
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
Figure 6.5 Scenario 4: Two OBSSs, one AP within range of two other
Scenario 4 represents the case of “neighborhood capture effect” [66] where there are three
BSSs, with one of them being in between the two others which do not hear to each other. As a
result, the central BSS suffers from a disproportionate degradation in throughput dependent
upon the total traffic in all three BSS. Hence, the two side networks may monopolize the
wireless medium, preventing the central one from getting any traffic through.
Figure 6.6 Scenario 5: Three OBSSs, two APs within range of each other
Scenario 5 depicts an overlapping scenario very similar to Scenario 3, with the only difference
being that just the two out of three APs are within the range of each other.
73
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
Figure 6.7 Scenario 6: Three OBSSs, APs not within range of each other, shared STAs
Scenario 6 illustrates an overlapping scenario of three OBSSs where the APs are out of range
of each other (hidden APs) but there are stations from all three BSSs that overlap.
Furthermore, there hidden nodes from each BSS.
Figure 6.8 Scenario 7: Two OBSSs, APs not within range of each other
Scenario 7 depicts an overlapping scheme similar to the scenario 4 with the “neighborhood
capture effect”, where one BSS is in between the other two. In this case, however, APs are not
within the range of each other but only some stations of each BSS are overlapping.
74
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
6.3 Importance of the OBSS Problem
It is expected that the number of the OBSSs in IEEE 802.11 aa/ac becomes larger than in
legacy standards (e.g. IEEE 802.11 a/n) because of both frequency bandwidth extension and
increased number of WLAN devices [67]. Furthermore, as mentioned earlier there are cases
where a station transmits in a 20 MHz channel, while also there are other cases where 40 MHz
channels are being used. 20 MHz frames are transmitted in 20 MHz channel as in normal,
while 40 MHz frames are transmitted occupying two adjacent 20 MHz channels. A major
issue may arise when 40 MHz stations coexist with 20 MHz stations in the same channel, due
to low possibility of having two 20 MHz channels vacant at the same time. Moreover, to
acquire both channels for 40 MHz communication, contention probability will increase to
double, as the 40 MHz stations will have to content with 20 MHz stations in both channels.
On the other hand, 20 MHz stations cannot understand those data sent in 40 MHz, thus they
may interfere with 40 MHz communication [68]. A solution to this problem is presented in
[68], where Y. Utsunomiya et al. suggested a MAC protocol to realize coexistence between 20
MHz and 40 MHz stations. In the proposed protocol, an AP controls the 20 MHz and 40 MHz
stations by setting a MAC level protection mechanism, to them, and divides the time into 20
MHz and 40 MHz periods.
In IEEE 802.11 aa/ac, 80 MHz of channel bandwidth is mandatory, thus there will only five
non-overlapping channels; 36-38, 52-64, 100-112, 116-128 and 149-161 plus a sixth, 132-144
with a regulatory change, as it is illustrated in Figure 6.9. In addition, there is the option of
160 MHz of channel bandwidth with only two non-overlapping channels; 36-64 and 100-128.
With this in mind, it becomes easily understandable that coexistence of different MHz stations
in the same channel, becomes even complicated.
Figure 6.9 The non-overlapping channels in the 5 GHz band [61]
75
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
In [69], the authors carried out an analysis of OBSS with the intention of determining the
criteria and features that will be used to provide a solution to the OBSS problem. Using
empirical propagation formula, they intent to estimate the maximum potential number of
overlapping networks for various residential scenarios. Table 6.4 illustrates the their results.
Detached Houses
12
Terraced Houses
16
Townhouses
25
Single Layout Apartments
28
Double Layout Apartments
53
Table 6.4 Maximum potential number of APs within range [69]
Figure 6.10 Apartment block single layout [69]
76
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
Figure 6.11 Non-overlapping channels in 2.4 GHz band for 20/40 MHz
Since there do not exist non-overlapping channels of 80 MHz bandwidth in the 2.4 GHz band,
the IEEE 802.11 aa/ac will operate in the 5 GHz band, a spectrum with less interference.
Channel selection can only partially reduce the problem as, even in the 5 GHz band with 19
non-overlapping channels, we could have situations of 2 or 3 overlapping BSSs. When the
BSSs are using the legacy channel access (Distributed Coordination Function) or the EDCA
function of IEEE 802.11e, the traffic competes fairly between the networks resulting in
reduced bandwidth in all of them.
The negative effect on the performance is more severe when the Quality of Service
enhancements of IEEE 802.11e are used, as these enhancements break down and cannot
support the performance demands of multimedia streams. When one of the BSSs uses EDCA
with admission control, the Access Point controls the admission of new streams, based on the
information it has about its own BSS, but it has no control and no information about the traffic
in the other overlapping networks. Therefore, it cannot guarantee the desired protection to the
admitted flows and ensure they are given the agreed Quality of Service. In the case where one
of the networks uses the HCCA function, in which the AP schedules traffic during a
Contention-Free Period and grants Transmit Opportunities (TXOP) to the stations, this
network is able to protect the bandwidth in its own network, reducing the available bandwidth
77
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
for the rest of the networks. However, when two or more networks use the HCCA function,
the AP may not be able to allocate time when it needs to, as it has to obey the TXOPs of other
networks. Again, this results in reduced bandwidth for the scheduled protected traffic and no
real protection for QoS restraints.
6.4 The IEEE 802.11aa Task Group Approach
The approach that was selected by the IEEE 802.11aa Task Group to manage the Overlapping
BSS situation is to provide a decentralized mechanism for neighboring APs to exchange
information about the QoS depended traffic load in each BSS. This information can be used
for more efficient channel selection and, if it is necessary for BSSs to share a channel, it
allows APs to cooperate and work as one bigger QoS-aware network, sharing fairly the
wireless medium. This approach is designed to mitigate the effects of overlapping BSSs and
provide the means to [70]:

Advertise QoS load information for channel selection.

Extend the admission control and scheduled mechanisms to a distributed environment.

Enable the coordination of scheduled HCCA TXOPs between HCs operating with
overlapping BSSs.
6.4.1
QLoad Report Element
The algorithm is based on the exchange of the QLoad Report Element between access points.
This Report Element is transmitted with one of the following ways:

When requested by another AP with a QLoad Request frame. For example, an AP
performing the channel selection procedure.
Figure 6.12 QLoad frame transmission, case 1
78
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa

Whenever there is a change in the values it includes. For example, when a new QoS
stream is accepted.
Figure 6.13 QLoad frame transmission, case 2

Optionally, when the dot11QLoadReportActivated is true, it is included in Beacons
every dot11QLoadReportIntervalDTIM.
Figure 6.14 QLoad frame transmission, case 3
The Report Element is exchanged directly between APs that are in direct range of each other,
as in Figure 6.2. However there is the case that a network overlaps with some stations in one
BSS but cannot be seen by the AP as illustrated in Figure 6.4. For these cases the AP can use
the neighbor report capability that was defined in IEEE 802.11k amendment. With this, it can
request from its associated stations to scan the medium for neighboring APs and send back a
Beacon Report, which may contain the QLoad Report Element.
Figure 6.15 QLoad Report Element format
79
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
The first three fields have the QLoad field format shown in Figure 6.16. The Mean field is the
mean medium time measured in units of 32μs per second for the sum of all the traffic streams,
and the standard deviation represents the statistical standard deviation from this mean. The
two other fields represent the number of AC_VO and AC_VI streams that are active and can
be used for some of the calculations. The way the mean and standard deviation is calculated is
included in the informative annex. It is based on the way medium time is calculated in the
recommended practices for admission control in the IEEE 802.11-2007 standard, but also
adds an overhead factor, based on the number of streams, to estimate the medium overhead
caused by the access time (AIFSN, backoff). The mean value for each traffic stream is
calculated based on its mean data rate, and the standard deviation is calculated by taking into
account the minimum and maximum data rate. When admission control is not used, the AP
has to monitor the data rate of QoS streams (those belonging in AC_VI and AC_VO access
categories) to calculate the mean and maximum medium time. These values for all the streams
are then summed to obtain the total mean and standard deviation.
Figure 6.16. QLoad Field format
6.4.2
Calculating field values
The Potential Traffic Self informs about the total QoS traffic an AP can expect. It is always
equal to or greater than the Allocated Traffic Self field. The potential traffic is calculated by
keeping an archive of the maximum value of any of the fields (mean, AC_VO streams,
AC_VI streams) over a seven day period. Any streams that get rejected by admission control
because of overallocation are also added to the value of the potential traffic. The Allocated
Traffic Self includes all the QoS traffic an AP has allocated at any given time. It changes
every time a new stream is accepted by EDCA admission control or the HCCA hybrid
controller, and every time such as stream is deleted. The Allocated Traffic Shared sums the
values of all the AP’s there is information about, including the self AP. It can be useful to
avoid the neighborhood capture effect. This way an AP knows that another AP is overallocated due to sharing with a neighboring AP that is out of its own range.
80
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
The HCCA Peak represents the total TXOP requirements of admitted HCCA traffic streams in
time units over one second and the HCCA Access factor is calculated as a fraction of the sum
of HCCA Peaks of overlapping networks plus the self network over one second. As this is a
sum over many networks, it may have a value greater than one, meaning that the total traffic is
greater than the capacity of the wireless medium. Then the EDCA Access factor is calculated
by adding the Potential Traffic of all overlapping networks including the self network, and
subtracting the HCCA traffic to obtain the EDCA traffic. Like the HCCA Access Factor it is
expressed as a fraction possibly greater than one.
The Overlap number represents the number of other APs sharing the same channel that this
AP has detected by detecting their Beacons. The sharing policy field shows if the AP
currently uses a static or dynamic sharing policy. A static sharing policy means that the
medium time of each AP does not change when the allocated traffic changes. In a dynamic
sharing policy the medium time of each AP may change as it allocates more traffic. The
optional subelements field is reserved for future versions and is not currently used.
The calculation of these fields, the channel selection procedure (sub-chapter 6.4.3) and the
sharing schemes (sub-chapter 6.4.4) are included in an informative annex which is intended
only as a suggestion and therefore it is not obligatory for an implementation.
6.4.3
Channel Selection
The channel selection procedure is as follows. The AP scans all the available channels for
Beacons by other APs and records for each channel the APs that use it, including their QoS
capabilities, and if they support the QLoad report, the information about overlaps and QoS
traffic. With these information the AP can select a channel in the following order:

A channel free from any overlaps with other APs

The channel with the least overlaps with other QoS capable APs

Depending on whether the AP uses EDCA admission control or a Hybrid Controller, it
should select the channel with the least overlaps with APs that use the same
transmission mode.

From these channels, it should select the one that reports the less Overlaps in its
QLoad report.

Finally, it should select the channel that shows the less potential traffic according to
the QLoad reports.
81
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
More specifically, the recommended method for channel selection can be implemented by
adoption of the following procedures [70]:

Create a list of the available channels. Typically this is the list of channels allowed by
regulation in the operating regulatory domain, however this list might be modified by
management policy (e.g. removing overlapping channels, avoiding radar detect
channels).

Create an array for each available channel that allows the recording of the QoS AP
count, Admission Control Mandatory count, HC count, overlap count and potential
load for that channel.

Step through the list of available channels, listening for beacons for at least
dot11OBSSScanPassiveTotalPerChannel TUs per channel.

Upon completion of the scan of a channel, process the beacons received on that
channel, filtered to the set of unique BSSIDs:
◦ Using the capabilities signaled in the beacon, modify the QoS AP count,
Admission Control Mandatory count, HC count, overlap count and potential load
of the channel array for the primary channel indicated in the received beacon.
◦ If the AP is using a channel bandwidth that is greater than the channel spacing (e.g.
when using the 2.4GHz band or when the overlapping AP allows 40MHz HT
PPDUs in its BSS) also update the channel array for channels that are affected by
this overlapping BSS. For example a beacon received on channel 2 indicating a
20MHz BSS also affects channels 1, 3 and 4.

Upon completion of scanning all of the channels, the AP has information on the
number of APs and the potential load of each channel, including co-channel BSSs.

If the channel array indicates that there are channels with no other APs, it is
recommended to randomly choose one of these “empty” channels.

Otherwise, create a list of candidate channels by selecting only the channels with the
lowest number of QoS APs. For example if the channel scan procedure indicated that
there were two QoS APs on channel 3, three QoS APs on channel 6 and two QoS APs
on channel 11, the list of candidate channels would contain 3 and 11.
◦ If this list contains one or more channels with non QoS APs, then filter the list for
the least number of APs.
◦ If this list contains more than one channel, it is recommended to randomly choose
one of these channels.
82
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa

If this list contains more than one channel and the AP has been configured to use
Admission Control Mandatory for AC_VI or AC_VO:
◦ Filter the list for the minimum count of QoS AP where the EDCA Parameter Set
element is present in the Beacon frame and with Admission Control Mandatory not
set for AC_VI or AC_VO.
◦ If this list contains more than one channel, filter the list for the minimum count of
QoS AP with Admission Control Mandatory set for AC_VI or AC_VO and that
does not indicate support for QLoad reporting.
◦ If this list contains more than one channel, filter the list for the minimum count of
QoS AP with an HC and that does not indicate support for QLoad reporting.
◦ If this list contains more than one channel, filter the list for the minimum count of
QoS AP with an HC, and that indicate support for QLoad reporting (as indicated
by the QLoad Report field equal to 1 in the Extended Capabilities element).
◦ If this list contains more than one channel, filter the list for the minimum count of
QoS AP with Admission Control Mandatory set for AC_VI or AC_VO and that
indicates support for QLoad reporting (as indicated by the QLoad Report field
equal to 1 in the Extended Capabilities element).

If this list contains more than one channel and the AP has an HC:
◦ Filter the list for the minimum count of QoS AP with an HC and that does not
indicate support for QLoad reporting.
◦ If this list contains more than one channel, filter the list for the minimum count of
QoS AP with Admission Control Mandatory set for AC_VI or AC_VO and that
does not indicate support for QLoad reporting.
◦ If this list contains more than one channel, filter the list for the minimum count of
QoS AP with an HC, and that indicate support for QLoad reporting (as indicated
by the QLoad Report field equal to 1 in the Extended Capabilities element).
◦ If this list contains more than one channel, filter the list for the minimum count of
QoS AP with Admission Control Mandatory set for AC_VI or AC_VO and that
indicates support for QLoad reporting (as indicated by the QLoad Report field
equal to 1 in the Extended Capabilities element).
◦ If this list contains more than one channel, filter the list for the minimum count of
QoS AP where the EDCA Parameter Set element is present in the Beacon frame
and with Admission Control Mandatory not set for AC_VI or AC_VO.
83
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa

If this list contains more than one channel, filter the list to the set of channels with the
minimum overlap count.

If this list contains more than one channel, filter the list to the set of channels with the
minimum potential load.

6.4.4
From the remaining channels in this list, randomly choose one of these channels.
Sharing in an OBSS situation
If the EDCA Access Factor is greater than one, then there is a potential over-allocation of the
WM. APs are advised to avoid this in the Channel selection process but if over-allocation
exists, then a sharing scheme is recommended to ensure that each AP has a fair share of the
bandwidth, but more importantly, to ensure that any already admitted or scheduled QoS
stream are not impaired by the addition of streams from any overlapping BSS. The EDCA
Access Factor, HCCA Access Factor and Potential Traffic Self fields in the QLoad Report are
provided to enable sharing schemes to be used.
The sharing scheme also protects an AP from the neighborhood effect where it has neighbors
that are hidden from each other. A major objective of an OBSS sharing scheme is that if a
QoS stream is allocated or scheduled, then it is not compromised by the addition of further
streams from any overlapping BSS that would cause the medium to be over-allocated. This is
achieved if the APs in overlapping BSSs cooperate [70].
The standard proposes two sharing schemes as examples of static and dynamic sharing
policies. The goal of both these schemes is to ensure that the total of QoS traffic admitted by
the overlapping networks does not exceed the capacity of the wireless medium. They are also
useful to ensure that the overlapping networks share fairly the available resources. Each AP
calculates a Maximum Allocation Value, based on the number of overlapping QoS APs, also
taking into account non-QoS APs that overlap in order to provide some resources for non-QoS
traffic, too.
The first scheme is called Proportional Sharing and it is an example of a static sharing policy.
Each AP calculates its maximum allowable traffic based on the EDCA and HCCA access
factors. When the AP wants to add a new traffic stream it checks if the sum will exceed its
maximum allowable traffic. The second scheme is called On-demand Sharing and is an
example of a dynamic sharing policy. The AP adds new streams based on the maximum
Allocated Traffic Self value from the overlapping APs.
84
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
It is suggested that APs use the On-demand Sharing scheme until they reach the Maximum
Allocation Value for any of the overlapping networks, and then use the proportional sharing
scheme for any subsequent requests.
6.4.4.1 Proportional Sharing scheme [70]
When using the proportional sharing scheme, the AP examines the sum of the EDCA and
HCCA Access Factors in the QLoad Reports from each overlapping BSS, including its own
QLoad, and determines the maximum. This maximum value is termed the “combined access
factor”.

If the maximum value from the combined access factor is less than or equal to MAV,
the AP is advised to allocate only up to its advertised Potential Traffic Self traffic.

If the maximum value from the combined access factor is greater than MAV, then the
AP is advised to allocate only up to a value of its Potential Traffic Self divided by the
combined access factor, multiplies by MAV.
In the proportional sharing scheme, before an AP allocates a new Medium Time or schedules
a new TXOP in response to an ADDTS Request, ti checks that this addition does not exceed
its sharing limit, as follows:

If the EDCA Access Factor is less than or equal to MAV, then the AP allocates up to
its advertised Potential Traffic Self, with the composite stream (MAX traffic)
calculated as:
MAXtraffic = μtot + 2 σtot.

If the EDCA Access Factor is greater than MAV, the AP carries out the following:
◦ Calculate the peak traffic value of the Potential Traffic Self, using:
Peak = μtot + 2 σtot.
◦ Divide this value by the combined access factor and multiply by MAV. This is
termed the maximum allowable Potential Traffic Self traffic.
◦ Calculate the resulting value of the Allocated Traffic Self if the new TSPEC is
accepted, as explained in aa.2.4, and then calculate the resulting peak value using:
Peak = μtot + 2 σtot.
85
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
◦ If the resulting peak value, calculated in step 3 is greater than the maximum
allowable Potential Traffic Self traffic, then the TS Request is advised to be
rejected.
◦ If the resulting peak value, calculated in step 3 is less than the maximum allowable
Potential Traffic Self traffic, and the TS Request is for EDCA admission, then it is
advised to be accepted.
◦ The AP then should check that it is possible to schedule TXOPs using the HCCA
TXOP advertisement.
If the new stream is allocated, then the AP updates the appropriate fields in its QLoad
element.
6.4.4.2 On-demand Sharing scheme [70]
The On-demand Sharing scheme is as follows:

Before allocating a new stream, the AP examines the Allocated Traffic Shared values
in the QLoad Reports from each overlapping BSS, including its own QLoad, and
selects the maximum Allocated Traffic Shared value which has the highest peak value,
using:
Peak = μtot + 2 σtot.
The AP also notes the number of AC_VI and AC_VO streams in this maximum Allocated
Traffic Shared Field.

The AP adds the requested new stream (new) to the selected maximum Allocated
Traffic Shared value (max) determined in step 1, using:
μ = μnew + Peak
σ = √σ2 new + σ2 max

The AP then calculates the peak value for the new composite stream calculated in step
2, using:
Peak = μ + 2σ
86
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa

Using the values of the AC_VI and AC_VO streams noted in step 1, plus the stream
represented be the new stream, the AP determines the new EDCA Access Factor and
then the combined access factor, as described in aa.4.2.2.

Multiply the peak value calculated in step 3 by the EDCA Access Factor, determined
in step 4. This is the new Peak Traffic requirement.

If this Peak Traffic requirement value calculated in step 5 is greater than MAV, then
the AP is advised to refuse to allocate the new stream.

If the peak value calculated in step 5 is less than or equal to MAV, and the new
allocation is for EDCA admission ADDTS, then the AP allocates that new traffic.

If the peak value calculated in step 4 is less than or equal to MAV, and the new
allocation is for HCCA ADDTS, the AP should check if it possible to schedule TXOPs
using the HCCA TXOP advertisement.
If the new stream is allocated, then the AP updates the appropriate fields in its QLoad
element.
6.4.5
HCCA TXOP Negotiation
In order to provide a means of synchronizing the TXOPs between overlapping networks that
use HCCA or mixed HCCA and EDCA mode, the IEEE 802.11aa standard provides a set of
action frames and information elements that can be exchanged between APs in direct contact
to inform each other of their TXOP schedules and allow them to schedule TXOP for newly
admitted traffic streams without compromising the schedule of other APs.
Collaboration candidates are APs that can exchange frames directly without the use of a third
station. The AP keeps a table of all its collaboration candidates. Collaboration candidates
support both protected and unprotected (public) TXOP negotiation. If peer APs have both
protected and unprotected TXOP negotiation activated, then the protected mode is used.
The TXOP schedule coordination is achieved with the use of HCCA TXOP Advertisement
and HCCA TXOP Response frames, which are represented in Figure 6.17 and Figure 6.18,
respectively. When the AP wants to schedule a TXOP for a new traffic stream, it sends a
TXOP advertisement frame to all the APs in the table with its current active TXOP
reservations and the proposed new reservation. An HCCA TXOP Count element is also
included in the Beacon frame to indicate that an HCCA TXOP schedule has changed.
87
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
Figure 6.17 HCCA TXOP Advertisement Frame
Figure 6.18 HCCA TXOP Response Frame
An AP shall not send an Add Traffic Stream (ADDTS) Response action frame to the
requesting station until one of the following conditions occurs [70]:

The AP has received an HCCA TXOP Response frame from all the APs to which
HCCA TXOP advertisements were sent, with the status field equal to SUCCESS.

At least two beacon frames have been received from all the APs to which HCCA
TXOP advertisements were sent.

A beacon containing the HCCA TXOP Update Count element is received from all the
APs to which HCCA TXOP advertisements were sent.

A period of three dot11BeaconPeriod TU has elapsed.
If an AP receives another TSPEC request while waiting for one of the above conditions to
occur, it shall delay processing this additional TSPEC request until one of the above
conditions occurs.
Upon reception of an unprotected HCCA TXOP Advertisement, an AP with
dot11PublicTXOPNegotiationActivated set to true, shall discard any entries that correspond to
the MAC address of the AP that sent the HCCA TXOP Advertisement and shall prepare a
response using the procedures below [70]:

An AP with dot11ProtectedTXOPNegotiationActivated set to false shall discard any
received Protected HCCA TXOP Advertisement frames.

An AP with dot11ProtectedTXOPNegotiationActivated set to true that does not have
an active security association with a peer AP that indicates support for protected
HCCA TXOP negotiation shall use the AP PeerKey Protocol and the Authenticated
Mesh Peering Exchange (both are defined in the original IEEE 802.11 standard) to
88
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
negotiate security parameters and secure the Protected HCCA TXOP Advertisement
frames.
Upon reception of a valid Protected HCCA TXOP Advertisement frame, an AP with
dot11ProtectedTXOPNegotiationActivated set to true, shall discard any entries that
correspond to the MAC address of the AP that sent the HCCA TXOP Advertisement and shall
prepare a response using the procedures below [70]:

If the HCCA TXOP Advertisement frame (either protected or unprotected) has not
been discarded due to the procedures above, the AP shall create a new entry for each
TXOP reservation in the Active TXOP Reservations field of the (Protected) HCCA
TXOP Advertisement frame.

If the HCCA TXOP Advertisement frame (either protected or unprotected) has not
been discarded due to the procedures above, the AP shall inspect its HCCA schedule
to check if the TXOP Reservations given in the Pending TXOP Reservations field of
the HCCA TXOP Advertisement frame is in conflict with an existing accepted HCCA
TXOP, allocated by itself. If there is no conflict, the AP shall send an HCCA TXOP
Response frame with the status field set to SUCCESS and create an entry for each
TXOP Reservation in the Pending TXOP Reservations field in the HCCA TXOP
Advertisement.

If the HCCA advertisement was sent using an unprotected Public Action frame, the
HCCA TXOP Response shall be send using an unprotected Public Action frame, too.

If the HCCA advertisement was sent using a Protected HCCA TXOP Advertisement
frame, the HCCA TXOP Response shall be sent using a Protected HCCA TXOP
Response frame.

If the AP detects that the TXOP given in the (Protected) HCCA TXOP Advertisement
frame is in conflict with an existing accepted HCCA TXOP and this AP is not itself in
the process of processing an ADDTS request, it shall send a (Protected) HCCA TXOP
Response frame with the status field set to TS_SCHEDULE_CONFLICT and the
Alternate Schedule field set to a period of time that does not conflict with any
currently accepted HCCA TXOPs and the Avoidance Request field absent. The
duration sub-field of the Alternate Schedule field should be greater than or equal to the
duration sub-field of the schedule field in the (Protected) HCCA TXOP Advertisement
frame. The duration sub-field of the Alternate Schedule field may be less than the
89
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
duration sub- field of the schedule field in the (Protected) HCCA TXOP
Advertisement frame, when there is an insufficient period of time that does not
conflict with currently accepted HCCA TXOPs.

If the AP detects that the TXOP given in the (Protected) HCCA TXOP Advertisement
frame is in conflict with an in-progress ADDTS request for a HCCA TXOP for which
HCCA TXOP Response frames have not been received, it shall send a (Protected)
HCCA
TXOP
Response
frame
with
the
status
field
set
to
TS_SCHEDULE_CONFLICT with the Alternate Schedule and Avoidance Request
fields set according to the following rules:
◦ If MIX(MACr) < MIX(MACi), the Alternate Schedule field is set to a value that
does not conflict with any accepted HCCA TXOPs and also does not conflict with
the TXOP of the in-progress ADDTS request. The Avoidance Request field is set
to the TXOP of the in-progress ADDTS request.
◦ If MIX(MACr) > MIX(MACi), the Alternate Schedule field is set to the value from
the TXOP Reservation from the TXOP Advertisement frame. The Avoidance
Request field is set to a time period that does not conflict with any accepted HCCA
TXOPs nor the TXOP in the Alternate Schedule field and has sufficient duration
and service interval to meet the requirements of the in-progress ADDTS request.
Where:
◦ MACr is the MAC address of the AP that received the TXOP Advertisement
frame.
◦ MACi is the MAC address of the AP that sent the TXOP Advertisement frame.
◦ The MIX function takes the 6 octets of a MAC address and computes a 6 octet
value:
MIX(MAC) = MAC[4] || MAC[5] || MAC[0] || MAC[1] || MAC[2] || MAC[3]
90
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
Case
Status Code
No conflict with
SUCCESS
Alternate Schedule Field Avoidance Request Field
Not present
Not present
existing or inprogress schedules
Conflicts with
TS_SCHEDUL Period of time that does
existing schedule,
E_CONFLICT not conflict with any
no ADDTS request
currently accepted HCCA
in progress
TXOPs
Not present
Conflict in-progress TS_SCHEDUL Period of time that does
Schedule of in-progress
schedules,
ADDTS request
E_CONFLICT not conflict with any
MIX(MACr) <
currently accepted HCCA
MIX(MACi)
TXOPs nor the inprogress ADDTS request
Conflict in-progress TS_SCHEDUL Same schedule that was in Period of time that does
schedules,
E_CONFLICT the TXOP Advertisement not conflict with any
MIX(MACr) >
currently accepted HCCA
MIX(MACi)
TXOPs nor the period
given in the Alternate
Schedule field
Table 6.5 Contents of the HCCA TXOP Response frame
According to [70], the AP shall also keep a record of the TXOP proposed in the alternate
schedule field in a TXOP avoidance record and should avoid scheduling any new HCCA
TXOPs in this proposed period until any of the following conditions occurs:
i. A period of dot11HCCATXOPBeaconTimeout multiplied by dot11BeaconPeriod TUs
has elapsed.
ii. The AP with dot11PublicTXOPNegotiationActivated set to true receives an
unprotected HCCA TXOP Advertisement Public Action frame from the AP to which
the unprotected HCCA TXOP Response frame was sent.
iii. The AP with dot11ProtectedTXOPNegotiationActivated set to true receives a
Protected HCCA TXOP Advertisement frame from the AP to which the Protected
HCCA TXOP Response frame was sent.
91
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
Finally, the standard also provides an algorithm for synchronizing the clocks between the
collaborating APs so the TXOP reservations can work properly. The APs constantly watch for
time drift between them and adjust to synchronize with the slower AP.
6.5 Related Work
The OBSS problem has concerned a lot of researchers who have tried to suggest various
solutions. Some of these proposals are categorized and are presented in following subchapters. The categorization has been made based on whether the work suggests single
channel management or dynamic channel switching. However, some of the proposals cannot
be grouped to any of these two categories, thus they are presented separately.
6.5.1
Single channel management
An approach to the OBSS problem is to create a mechanism where the overlapped BSSs are
able to manage access in the shared channel, in a way that would be fair and would prevent
interference. Schemes that are based on this approach, require from the APs from each
overlapped BSS to periodically communicate with each other, in order to cooperatively
coordinate their channel access operations. This means that such a scheme is only applicable
in an OBSS situation where the APs are in direct range with each other, such as in the
topologies described in Figure 6.2 and Figure 6.4.
Bo Han et al. [62], proposed the use of Channel Access Throttling (CAT) in order to manage
radio resources for overlapping BSSs. With CAT an AP of each BSS is provided with a
coordination mechanism in order to control channel access parameters of its member stations
on the fly. With this mechanism, privileged channel access can be supported for a particular
BSS at a particular time. This can be achieved, for example, by assigning higher priority
channel access parameters to stations associated with this BSS. Furthermore, by tracking how
much each BSS has been given privileged channel access, it is possible to achieve
proportional partitioning of channel capacity among the overlapped BSSs, thus improve
channel utilization. In their scenario, the authors present a simple two-priority-group model
which, however, can be extended to more complex models. For instance, a model that allows
more than one BSSs to be assigned into the high-priority group, at the same time.
A similar solution is presented in [71]. The author combines two distributed schemes to
achieve throughput fairness among overlapped BSSs. The first one uses Dynamic Network
92
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
Allocation Vectors (DNAVs) which dynamically adjust the channel interference by
temporarily stopping transmissions in some BSSs, while the second scheme uses Forced
Hand-Offs (FHO) to force stations, placed within the overlapped area, to hand off. The
DNAV is used for each AP to calculate its BSS weighted throughput value and serves to keep
throughput fairness among the BSSs. If an AP finds that its own weighted value is higher than
the average of all the received weighted throughput values by a certain amount, it increases
the NAV duration in its own BSS; otherwise the NAV duration is decreased. Without using
NAV to control the inter-BSS interference, BSSs having a lower weighted value can force
some associated stations to hand off to neighboring BSSs that have higher weighted
throughput value.
In [72], a spectrum sharing concept is proposed that allows BSSs to coexist in a single
channel, while providing QoS at the same time. With this scheme the overlapped BSSs
multiplex their channel access in the time domain, where each system leaves some channel
capacity that can be used as idle period for other systems that may want to compete for
medium access. In a way this method can be considered as “TDMA between systems”.
However, distributing idle period in a random manner may increase the collisions. Thus, the
proposed scheme is implemented in a way that it will occupy the channel in a way keeping
idle periods in a regular pattern. Therefore, this scheme can be referred as Regular Channel
Access (RCA). RCA is helpful to the OBSSs in two ways: first it allows them to reliably
predict the length of the idle periods and their starting offset, and second, it helps each BSS to
utilize the idle periods for its own transmissions, reducing to the minimum the collisions
caused by orthogonality in time. The simulation results have shown that the proposed solution
can achieve a fair throughput sharing, keep the delay of multimedia streams transmission in
low levels and reduce the total interference and overhead.
6.5.2
Dynamic channel switching
Another approach to the problem is channel switching after OBSS detection. However,
recognizing the existence of OBSS is not as simple as it may sounds. First of all, stations do
not always receive frames correctly from OBSSs, thus they may not be able to check the
BSSID. This is even worse as the traffic becomes heavier and the transmission rates of frames
originating from the OBSSs are higher. Another difficulty is that if the cell radius of the
domestic BSS is smaller that that of the OBSS, some stations in the OBSS may transmit
frames asynchronously and interfere with transmissions in the domestic BSS. If this
93
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
interference causes frame errors, the domestic stations may assume that this interference is
caused by degradation of channel condition [65].
With that in consideration, [65] proposes a method where stations can detect packets
corrupted by interference, even while they receive frames from the domestic BSS. Stations
consider two examinations in order to detect interferential packets. The first is to determine
whether differential of averaged Received Signal Strength Indicator (RSSI) exceeds threshold.
The authors explain that if both the RSSI exceeds two threshold values (one positive and one
negative) and the Cyclic Redundancy Check (CRC) indicates that the packet is received
incorrectly, then it is decided that interferential packet is detected. The second examination is
to determine whether the frame duration derived from header information equals the frame
duration as measured by the RSSI. Finally, if a station detects interferential packets with any
of the described methods, it reports the existence of another BSS to the AP which, in turn,
announces channel switching to all the associated stations in the domestic BSS. Simulation
results have shown that this mechanism has a high probability of detecting interferential
packets and, also, that throughput can be kept by channel switching.
[64] describes the implementation of a wireless mesh network that supports Dynamic Channel
Switching (DCS). Although DCS enables wireless networks to avoid frequency channels with
serious interference, it significantly increases the complexity of the networks' routing
protocol. The authors integrated the proposed mechanism with fault-tolerant and loadbalancing routing to demonstrate that it is still practical to apply DCS to IEEE 802.11-based
APs. The resulted wireless mesh network is called “Carlsbad”. The simulation results have
shown that despite the additional overhead, DCS can improve the overall throughput of a
wireless network, for both TCP and UDP.
6.5.3
Other solutions
In [73], the authors propose an alternative method for nodes in a wireless network to detect
and ignore frames received from a BSS, other than the one they are associated with. To
achieve that, a new field is added in the Physical Layer Convergence Protocol (PLCP) header
(Figure 6.19). This field contains a portion of BSSID, among others, to indicate who must
decode this frame. Upon reception of such a frame a node determines whether or not this field
is valid and, if it is, the node decodes the succeeding MAC frame. Otherwise, the node drops
the MAC frame without decoding it and returns to medium carrier sensing mode. Even though
94
Overlapping Basic Service Set (OBSS) Issues and IEEE 802.11aa
this solution improves the area throughput in an OBSS environment, it does not provide a
complete solution since it lacks in several ways. For example, as mentioned earlier,
interference caused by OBSS may alter the transmitted data, thus the receiving nodes would
not be able to check the BSSID field of the PLCP header. Also, since PLCP is used only by
OFDM-based wireless LANs, this solution is limited only to this type of networks.
Figure 6.19 PLCP header modification
95
Chapter 7
Conclusions and Further Research
7.1 Conclusions
The current thesis focused on the research efforts that have been made by the IEEE 802.11aa
Task Group and aim at providing improvements for audio/video transmissions over wireless
LANs. To succeed this, the TGaa has defined various enhancements on the MAC layer. The
latest IEEE 802.11aa draft release, is the Draft 9.0 and was published in January 2012.
Initially, Chapter 1 of the current thesis presented a short description of the IEEE organization
and its family of protocols that operate on the wireless medium. A separate, more detailed
overview of the IEEE 802.11 protocols was given in Chapter 2, since our objective, which
was IEEE 802.11aa, belongs to this group of protocols. Chapter 3 contained a significant part
of this thesis which has been devoted in describing the main features and functionality of the
current IEEE 802.11 MAC layer, especially since the TGaa focuses on enhancing this
particular layer.
Thereafter, the main fields of study of the TGaa were categorized in four groups: the Stream
Classification Service, Interworking with 802.1AVB, Multicast and, finally, the OBSS issue.
The first two are closely related to each other, thus, they were presented together in Chapter 4,
whereas, the efforts about Multicast and the OBSS issues, were presented separately in
Chapter 5 and Chapter 6, respectively. Along with the presentation of the TGaa efforts, the
thesis also provided related approaches and proposals from individual researchers, who tried
to provide solutions on these particular fields, mostly before the formation of the task group.
Regarding the multicast issue, the individual proposals were further categorized based on the
OSI/ISO reference model layer that they apply on. An extra category has been created, the so
called Cross-Layer Solutions, where proposals that rely on more than one layers were
presented. Finally, the individual solutions for the OBSS issue were divided in two categories,
based on whether they suggest single channel management or dynamic channel switching.
7.2 Further Research
For future work, we are interested in creating an updated summary of the modifications and
the enhancements on the MAC layer, as defined in the first oncoming IEEE 802.11aa
96
Conclusions and Further Research
standard. It would be also interesting to evaluate its performance both in simulation, but
mostly in real conditions, and compare the results with those of the performance of legacy
IEEE 802.11 protocols.
97
List of Figures
Figure 2.1 IEEE 802.11 standards mapped to the OSI reference model [9]...............................8
Figure 2.2 Immediate and delayed block ACK sessions [15]...................................................13
Figure 2.3 MIMO with TXN transmit and RXN receive antennas [14]...................................17
Figure 3.1 MAC architecture [9]...............................................................................................24
Figure 3.2 Distributed Coordination Function [9].....................................................................25
Figure 3.3 CW incrementation example [9]..............................................................................27
Figure 3.4 Backoff Procedure for DCF [9]...............................................................................28
Figure 3.5 RTS/CTS/data/ACK and NAV setting [9]...............................................................28
Figure 3.6 Simplified algorithm of CSMA/CA.........................................................................29
Figure 3.7 Point Coordination Function [9]..............................................................................30
Figure 3.8 Reference implementation model [22].....................................................................32
Figure 4.1 EDCA operation with Intra-Access Category Prioritization [9]..............................36
Figure 4.2 Intra-Access Category Priority Element..................................................................37
Figure 4.3 Overview of IEEE 802.1 AVB................................................................................40
Figure 4.4 AVB Connections....................................................................................................41
Figure 4.5 Clocking hierarchy...................................................................................................42
Figure 4.6 Insertion of 802.1Q Tag in an Ethernet frame.........................................................42
Figure 5.1 Primary idea of BMMM..........................................................................................48
Figure 5.2 RAK frame format...................................................................................................49
Figure 5.3 BMW vs BMMM.....................................................................................................49
Figure 5.4 Comparison of the initial transmission phase..........................................................50
Figure 5.5 Comparison of the retransmission phase.................................................................50
Figure 5.6 RM and FB frame formats.......................................................................................51
Figure 5.7 An 802.11n A-MPDU based scheme.......................................................................52
Figure 5.8 ALC frame format....................................................................................................53
Figure 5.9 DirCast operational overview..................................................................................53
Figure 5.10 GP protocol network diagram................................................................................54
Figure 5.11 GP protocol over time diagram..............................................................................55
Figure 5.12 Each node consists of two layers...........................................................................56
Figure 5.13. Normal flooding (A) vs proposed scheme's flooding (B).....................................57
Figure 5.14. Control frames formats.........................................................................................58
98
List of Figures
Figure 5.15 Group Address Transmission Service with different policies..............................65
Figure 5.16 The proposed NACK based LBP...........................................................................66
Figure 5.17 Multicast Collision-Free Mechanism.....................................................................66
Figure 5.18 Multicast channel access: (a) Standard, (b) MCF..................................................67
Figure 6.1 Channels of 22 MHz bandwidth in the 2.4 Ghz band..............................................69
Figure 6.2 Scenario 1: Two OBSSs, APs within range of each other.......................................71
Figure 6.3 Scenario 2: Two OBSSs, APs not within range of each other.................................72
Figure 6.4 Scenario 3: Three OBSSs, APs within range of each other.....................................72
Figure 6.5 Scenario 4: Two OBSSs, one AP within range of two other...................................73
Figure 6.6 Scenario 5: Three OBSSs, two APs within range of each other..............................73
Figure 6.7 Scenario 6: Three OBSSs, APs not within range of each other, shared STAs........74
Figure 6.8 Scenario 7: Two OBSSs, APs not within range of each other.................................74
Figure 6.9 The non-overlapping channels in the 5 GHz band [61]..........................................75
Figure 6.10 Apartment block single layout [69].......................................................................76
Figure 6.11 Non-overlapping channels in 2.4 GHz band for 20/40 MHz.................................77
Figure 6.12 QLoad frame transmission, case 1.........................................................................78
Figure 6.13 QLoad frame transmission, case 2.........................................................................79
Figure 6.14 QLoad frame transmission, case 3.........................................................................79
Figure 6.15 QLoad Report Element format...............................................................................79
Figure 6.16. QLoad Field format...............................................................................................80
Figure 6.17 HCCA TXOP Advertisement Frame.....................................................................88
Figure 6.18 HCCA TXOP Response Frame.............................................................................88
Figure 6.19 PLCP header modification.....................................................................................95
99
List of Tables
Table 2.1 Published amendments [8]........................................................................................21
Table 2.2 Amendments in process............................................................................................23
Table 3.1 IEEE 802.11 Carrier Sensing Mechanisms...............................................................25
Table 3.2 UP-to-AC mappings [9]............................................................................................31
Table 3.3 Default EDCA Parameter Set element parameter values [9]....................................33
Table 4.1 UP-to-AC mappings with 6 transmitting queues.....................................................37
Table 5.1 Video rate adaptation based on PER for different transmission rates.......................61
Table 5.2 Characteristics Of Group Addressed Transmission Service Policies........................65
Table 6.1 WLAN channel frequencies in 2.4 GHz and their availability per region................69
Table 6.2 WLAN channel frequencies in 5 GHz and their availability per region...................70
Table 6.3 Radio bands that common IEEE protocols operate in...............................................70
Table 6.4 Maximum potential number of APs within range [69].............................................76
Table 6.5 Contents of the HCCA TXOP Response frame........................................................91
100
Bibliography
[1] B.H. Walke, S. Mangold, L. Brlemann, “IEEE 802 Wireless Systems: Protocols, MultiHop Mesh/Relaying, Performance and Spectrum Coexistence”, Wiley, January 2007.
[2] “IEEE 802.15 WPAN Task Group 3 (TG3)”, IEEE-802, [Online]. Available:
http://www.ieee802.org/15/pub/TG3.html (Last Accessed: April 2012).
[3] “IEEE 802.15 WPAN Task Group 5 (TG5) Mesh Networking”, IEEE-802, [Online].
Available: http://www.ieee802.org/15/pub/TG5.html (Last Accessed: April 2012).
[4] “IEEE 802.15 WPAN Task Group 6 (TG6) Body Area Networks”, IEEE-802, [Online].
Available: http://www.ieee802.org/15/pub/TG6.html (Last Accessed: April 2012).
[5] “IEEE 802.18 & 802.19”, IEEE-802, [Online]. Available: http://www.ieee802.org/miscdocs/GlobeCom2009/IEEE_802d18_and_d19_Kraemer.pdf (Last Accessed: April 2012).
[6] D.T. Wong, P.Y. Kong, Y.C. Liang, K.C. Chua, “Wireless Broadband Networks”, Wiley,
April 2009.
[7] “IEEE 802.22 Working Group on Wireless Regional Area Networks”, IEEE-802,
[Online]. Available: http://www.ieee802.org/22/ (Last Accessed: April 2012).
[8] G.R. Hiertz, D. Denteneer , L. Stibor, Y. Zang, X.P. Costa, B. Walke, “The IEEE 802.11
Universe”, IEEE Communications Magazine, vol. 48, issue 1, pp. 62-70, January 2010.
[9] IEEE Standard 802.11-2007, Wireless LAN Medium Access Control (MAC) and Physical
Layer (PHY) Specifications, June 2007.
[10] R.T. Valadas, A.R. Tavares, A.M. de Oliveira Duarte, A.C. Moreira, C.T. Lomba, “The
infrared physical layer of the IEEE 802.11 standard or wireless local area networks”, IEEE
Communications Magazine, vol. 36, issue 12, pp. 107-112, December 1998.
[11] “IEEE 802.11 Frequency Hopping Spread Spectrum”, Wireless-Center, [Online].
Available: http://www.wireless-center.net (Last accessed: April 2012).
[12] N.R. Prasad, “IEEE 802.11 system design”, In Proceedings of the IEEE International
Conference on Personal Wireless Communications (ICPWC), pp. 490-494, Nieuwegein,
Netherlands, 17-20 December 2000.
[13] L. Litwin, M. Pugel, “The principles of OFDM”, www.rfdesign.com, January 2001.
101
Bibliography
[14] B. O'Hara, A. Petrick, “IEEE 802.11 handbook: a designer's companion”, Wiley-IEEE
Press, January 2005.
[15] E. Perahia, R. Stacey, “Next generation Wireless LANs: Throughput, Robustness and
Reliability in 802.11n”, Cambridge University Press, 2008.
[16] S. Mangold, L. Berlemann, “IEEE 802.11k: improving confidence in radio resource
measurements”, In Proceedings of the IEEE 16th International Symposium on Personal,
Indoor and Mobile Radio Communications (PIMRC), vol. 2, pp. 1009-1013, Berne,
Switzerland, 11-14 September 2005.
[17] A.A. Tabassam, H. Trsek, S. Heiss, J. Jasperneite, “Fast and seamless handover for
secure mobile industrial applications with 802.11r”, In Proceedings of the IEEE 34th
Conference on Local Computer Networks (LCN), pp. 750-757, Lemgo, Germany, 20-23
October 2009.
[18] R.V. Nee, “Breaking the gigabit-per-second barrier with 802.11ac”, IEEE Wireless
Communications, vol. 18, issue 2, pp. 4, April 2011.
[19] M. Cast, “802.11 Wireless Networks: The Definitive Guide”, O'Reilly Media, April
2005.
[20] F.Y. Hung, I. Marsic, “Effectiveness of Physical and Virtual Carrier Sensing in IEEE
802.11 Wireless Ad Hoc Networks”, In Proceedings of the IEEE Wireless Communications
and Networking Conference (WCNC), pp. 143-147, Kowloon, China, 11-15 March 2007.
[21]
A.
vonNagy,
“Understanding
Wi-Fi
carrier
sense”
,
[Online].
http://revolutionwifi.blogspot.com/2011/03/understanding-wi-fi-carrier-sense.html
Available:
(Last
accessed: April 2012).
[22] D. Gao, J. Cai, K.N. Ngan, “Admission control in IEEE 802.11e wireless LANs”, IEEE
Network, vol. 19, issue 4, pp. 6-13, July-August 2005.
[23] F. Cali, M. Conti, E. Gregori, “Dynamic tuning of the IEEE 802.11 protocol to achieve a
theoretical throughput limit”, IEEE/ACM Transactions on Networking, vol. 8, issue 6, pp.
785-799, December 2000.
102
Bibliography
[24] G. Bianchi, I. Tinnirello, “Kalman filter estimation of the number of competing terminals
in an IEEE 802.11 network”, In Proceedings of the 22nd Annual Joint Conference of the IEEE
Computer and Communications, IEEE Societies, INFOCOM, vol. 2, pp. 844-852, 30 March-3
April 2003.
[25] M. Hui, L. Xing, L. Hewu, Z. Peiyun, L. Shixin, Y. Cong, “Dynamic optimization of
IEEE 802.11 CSMA/CA based on the number of competing stations”, In Proceedings of the
2004 IEEE International Conference on Communications, vol. 1, pp. 191-195, 20-24 June
2004.
[26] N. Qiang, I. Aad, C. Barakat, T. Turletti, “Modeling and analysis of slow CW decrease
for IEEE 802.11 WLAN”, In Proceedings of the 14th IEEE Conference on Personal, Indoor
and Mobile Radio Communications (PIMRC), vol.2, pp. 1717-1721, 7-10 September 2003.
[27] L. Wanming, Y. Baoping, L. Xiaoxong, M. Wei, “An enhanced service differentiation
mechanism for QoS provisioning in IEEE 802.11e wireless networks”, In Proceedings of the
10th International Conference on Advanced Communication Technology (ICACT), vol. 1, pp.
175-180, Gangwon, South Korea, 17-20 February 2008.
[28] “IEEE 802.1”, IEEE-802, [Online]. Available: http://www.ieee802.org/1/ (Last
Accessed: April 2012).
[29] G.M. Garner, R. Hyunsurk, “Synchronization of Audio/Video Bridging Networks Using
IEEE 802.1AS”, IEEE Communications Magazine, vol. 49, issue 2, pp. 140-147, February
2011.
[30] “CCNA Exploration: LAN Switching and Wireless”, Cisco Systems, 2007-2009, Chapter
3.
[31] G. Venkatesan, “Multimedia Streaming over 802.11 links”, IEEE Wireless
Communications Magazine, vol. 17, issue 2, pp. 4-5, April 2010.
[32] H. Zinner, J. Noebauer, K. Seitz, T. Waas, “A Comparison of Time Synchronization in
AVB and FlexRay in-vehicle networks”, In Proceedings of the 9th Workshop on Intelligent
Solutions in Embedded Systems (WISES), pp. 67-72, Regensburg, Germany, 7-8 July 2011.
[33] G.M. Garner, G. Feifei, K. den Hollander, J. Kongkyu, K. Byungsuk, L. Byoung-Joon, J.
Tae-Chul, J. Jinoo, “IEEE 802.1 AVB and its application in carrier-grade Ethernet”, IEEE
Communications Magazine, vol. 45, issue 12, pp. 126-134, December 2007.
103
Bibliography
[34] M. Jakovljevic, A. Ademaj, “Ethernet protocol services for critical embedded systems
applications”, In Proceedings of the IEEE/AIAA 29th Digital Avionics Systems Conference
(DASC), pp. 5.B.3-1–5.B.3-10, Salt Lake City, USA, 3-7 October 2010.
[35] J. Kuri, S.K. Kasera, "Reliable multicast in multi-access wireless LANs", In Proceedings
of the 18th Annual Joint Conference of the IEEE Computer and Communications Societies,
INFOCOM, vol. 2, pp. 760-767, New York, USA, 21-25 March 1999.
[36] W. Xiaoli, W. Lan, W. Yingjie, Z. Yongsheng, "A reliable and efficient MAC layer
multicast protocol in wireless LANs", In Proceedings of the IEEE 69th Vehicular Technology
Conference (VTC), pp. 1-5, Barcelona, Spain, 26-29 April 2009.
[37] K. Tang, M. Gerla, "MAC reliable broadcast in ad hoc networks", In Proceedings of the
Military Communications Conference (MILCOM). Communications for Network-Centric
Operations: Creating the Information Force. IEEE, vol. 2, pp. 1008-1013, 2001.
[38] S. Min-Te, H. Lifei, A. Arora, L, Ten-Hwang, "Reliable MAC layer multicast in IEEE
802.11 wireless networks", In Proceedings of the International Conference on Parallel
Processing, pp. 527-536, 18-21 August 2002.
[39] A. Lyakhov, V. Vishnevsky, M. Yakimov, "Multicast QoS support in IEEE 802.11
WLANs", In Proceedings of the IEEE International Conference on Mobile Adhoc and Sensor
Systems (MASS), pp. 1-3, Pisa, Italy, 8-11 October 2007.
[40] V. Srinivas, R. Lu, "An Efficient Reliable Multicast Protocol for 802.11-based Wireless
LANs", In Proceedings of the IEEE International Symposium on World of Wireless, Mobile
and Multimedia Networks & Workshops (WoWMoM), pp. 1-6, Kos, Greece, 15-19 June
2009.
[41] S.K.S. Gupta, V. Shankar, S. Lalwani, "Reliable multicast MAC protocol for wireless
LANs", In Proceedings of the IEEE International Conference on Communications (ICC), vol.
1, pp. 93-97, 11-15 May 2003.
[42] J. Wenbin, J. Hai, L. Xiaofei, Y. Zhi, "A multicast based bandwidth saving approach for
wireless live streaming system", International Journal of Smart Home, Security Engineering
Research Support, vol. 2, pp. 65-80, January 2008.
[43] M. Samokhina, K. Moklyuk, S. Choi, J. Heo, "Raptor code-based video multicast over
IEEE 802.11 WLAN", Seoul National University, June 2008.
104
Bibliography
[44] H. Fujisawa, K. Aoki, M. Yamamoto, Y. Fujita, "Estimation of multicast packet loss
characteristic due to collision and loss recovery using FEC on distributed infrastructure
wireless LANs", In Proceedings of the IEEE Wireless Communications and Networking
Conference (WCNC), vol. 1, pp. 399-404, 21-25 March 2004.
[45] M. Kappes, "An application-layer approach for multihop communication in IEEE 802 11
ad-hoc networks", In Proceedings of the IEEE Wireless Communications and Networking
Conference (WCNC), vol. 3, pp. 1359-1364, 21-25 March 2004.
[46] S. Karanth, T. Moscibroda, V. Navda, J. Padhye, R. Ramjee, L. Ravindranath, "DirCast
A practical and efficient Wi-Fi multicast system", In Proceedings of the 17th IEEE
International Conference on Network Protocols (ICNP), pp. 161-170, Princeton, USA, 13-16
October 2009.
[47] C. Bravo, A. Gonzalez, "Evaluation and improvement of multicast service in 802 11b",
Federico Santa Maria University, 2006.
[48] L. Zhi, G. Cheung, V. Velisavljevic, E. Ekmekcioglu, J. Yusheng, "Joint source-channel
coding for WWAN multiview video multicast with cooperative peer-to-peer repair", In
Proceedings of the 18th International Packet Video Workshop (PV), pp. 110-117, HongKong, China, 13-14 December 2010.
[49] X. Zhu, T. Schierl, T. Wiegand, B. Girod, "Video Multicast over wireless mesh nerworks
with scalable video coding (SVC)", Visual Communications and Image Processing, January
2008.
[50] M. Gerla, P. Zhao, B. Daneshrad, G. Pei, J.H. Kim, "MIMO-Cast: A Cross-Layer ad hoc
multicast protocol using MIMO radios", In Proceedings of the IEEE Military
Communications Conference (MILCOM), pp. 1-6, Orlando, USA, 29-31 October 2007.
[51] C.C. Hu, "Efficient cross-layer protocol for bandwidth-satisfied multicast in multi-rate
MANETs", Wireless Networks, vol. 17, issue 3, April 2011.
[52] J. Villalon, P. Cuenca, L. Orozco-Barbosa, Y. Seok, T. Turletti, "ARSM: a cross-layer
auto rate selection multicast mechanism for multi-rate wireless LANs", IET Communications,
vol. 1, issue 5, pp. 893-902, October 2007.
105
Bibliography
[53] J. Villalon, P. Cuenca, L. Orozco-Barbosa, Y. Seok, T. Turletti, "Cross-Layer
Architecture for Adaptive Video Multicast Streaming Over Multirate Wireless LANs", EEE
Journal on Selected Areas in Communications, vol. 25, issue 4, pp. 699-711, May 2007.
[54] W.S. Lim, D.W. Kim, Y.J. Suh, "Design of Efficient Multicast Protocol for IEEE
802.11n WLANs and Cross-Layer Optimization for Scalable Video Streaming", IEEE
Transactions on Mobile Computing, vol. 11, issue 5, pp. 780-792, May 2012.
[55] O. Alay, C. Li, A. Rai, T. Korakis, W. Yao, S. Panwary, "Dynamic Rate and FEC
Adaptation for Video Multicast in Multi-rate Wireless Networks", In Proceedings of the 5th
International Conference on Testbeds and Research Infrastructures for the Development of
Networks & Communities and Workshops (TridentCom), pp. 1-8, Washington, USA, 6-8
April 2009.
[56] K. Maraslis, P. Chatzimisios, A.C. Boucouvalas, "IEEE 802.11aa: Improvements on
video transmission over wireless LANs", In Proceedings of the IEEE International Conference
on Communications (ICC 2012), Ottawa, Canada, 10-15 June 2012.
[57] R. Suri, B. Hart, "Proposal for Addressing MRG Block Ack related comments", Doc.
IEEE 802.11-10/1179r0, September 2009.
[58] J. Miroll, Z. Li, T. Herfet, "Wireless feedback cancellation for leader-based MAC layer
multicast protocols: Measurement and simulation results on the feasibility of leader-based
MAC protocols using feedback cancellation on the 802.11aa wireless multicast network", In
Proceedings of the IEEE 14th International Symposium on Consumer Electronics (ISCE), pp.
1-6, Braunschweig, Germany, 7-10 June 2010.
[59] M.A. Santos, J. Villalon, L. Orozco-Barbosa, "Multicast Collision Free (MCF)
mechanism over IEEE 802.11 WLANs", In Proceedings of the 3rd Joint IFIP. Wireless and
Mobile Networking Conference (WMNC), pp. 1-6, Budapest, Hungary, 13-15 October 2010.
[60]
“ISM
band
–
technical
definition”,
YourDictionary,
[Online],
Available:
http://computer.yourdictionary.com/ism-band (Last accessed: April 2012).
[61] F. Yue, G. Daqing, A.B. McDonald, Z. Jinyun, “A two-level carrier sensing mechanism
for overlapping BSS problem in WLAN”, In Proceedings of the 14th IEEE Workshop on
Local and Metropolitan Area Networks (LANMAN), Crete, Greece, 18 September 2005.
106
Bibliography
[62] H. Bo, J. Lusheng, L. Seungjoon, R.R. Miller, B. Bhattacharjee, “Channel access
throttling for overlapping BSS management”, In Proceedings of the IEEE International
Conference on Communications (ICC), pp. 1-6, Dresden, Germany, 14-18 June 2009.
[63] K. Heck, “Wireless LAN performance in overlapping cells”, In Proceedings of the 58th
IEEE Vehicular Technology Conference (VTC), vol. 5, pp. 2895-2900, 6-9 October 2003.
[64] W. Gang, S. Sathyanarayana, C. Tzi-cker, “Implementation of dynamic channel
switching on IEEE 802.11-based wireless mesh networks”, In Proceedings of the 4th Annual
International Conference on Wireless Internet (WICON), Maui, USA, 17-19 November 2008.
[65] T. Tandai, K. Toshimitsu, T. Sakamoto, “Interferential packet detection scheme for a
solution to overlapping BSS issues in IEEE 802.11 WLANs”, In Proceedings of the IEEE
17th International Symposium on Personal, Indoor and Mobile Radio Communications
(PIMRC), pp. 1-6, Helsinki, Finland, 11-14 September 2006.
[66] D.D. Vergados, D.J. Vergados, “Synchronization of multiple access points in the IEEE
802.11 point coordination functions”, In Proceedings of the IEEE 60th Vehicular Technology
Conference, vol. 2, pp. 1073-1077, 26-29 September 2004.
[67] Y. Asai, T. Ichikawa, R. Kudo, K. Ishihara, T. Murakami, Y. Takatori, M. Mizoguchi,
“Interference management using beamforming technique in OBSS environment”, Doc. IEEE
802.11-10/0585r2, May 2010.
[68] Y. Utsunomiya, T. Tandai, T. Adachi, M. Takagi, “A MAC protocol for coexistence
between 20/40 MHz STAs for high throughput WLAN”, In Proceedings of the IEEE 63th
Vehicular Technology Conference (VTC), vol. 3, pp. 1136-1140, Melbourne, Australia, 7-10
May 2006.
[69] G. Smith, “Overlapping BSS analysis of channel requirements”, Doc. IEEE 802.1108/1470-04-00aa, February 2009.
[70] IEEE P802.11aa/D9.0, Wireless LAN Medium Access Control (MAC) and Physical
Layer (PHY) specifications – Amendment 2: MAC Enhancements for Robust Audio Video
Streaming, January 2012.
[71] Z. Dongmei, “Throughput fairness in infrastructure-based IEEE 802.11 mesh networks”,
IEEE Transactions on Vehicular Technology, vol. 56, issue 5, pp. 3210-3219, September
2007.
107
Bibliography
[72] M.M. Siddique, B.L. Wenning, A. Timm-Giel, C. Gorg, M. Muhleisen, “Generic
spectrum sharing method applied to IEEE 802.11e WLANs”, In Proceedings of the 6th
Advanced International Conference on Telecommunications (AICT), pp. 57-63, Barcelona,
Spain, 9-15 May 2010.
[73] K. Yano, S. Tsukamoto, M. Taromaru, M. Ueba, “Area throughput enhancement of
OFDM-based wireless LAN in OBSS environment by physical header modification and
adaptive array antenna”, In Proceedings of the IEEE 20th International Symposium on
Personal, Indoor and Mobile Communications (PIMRC), pp. 3059-3063, Tokyo, Japan, 13-16
September 2009.
108